You are on page 1of 123

LESSON 1

MATHEMATICAL MODELLING FOR MANAGEMENT DECISION


MAKING

Mathematics is the science of magnitude and number, and of all their relationships.
Mathematics is being applied in almost all fields. Management is not exception. In fact
there is a kind of mathematicism, that is, everything can be described in terms i.e.,
numbers and symbols. True.

1.1 MEASUREMENT AND MATHEMATICS

Yes, the world is moving towards becoming more and more quantitative. To be
quantitative is to be more precise and pointed in what is being said or done.
Measurement is fundamental aspect of being quantitative. To measure means, to express
the physical, biological, sociological and other observed phenomena in terms of numbers
or symbol such that the phenomena measured bear the same relationship as to the used
numbers or symbols have, inter se in terms of magnitude and other relations.
The world of business has many phenomena and all these have to be measured for
their management and manipulation. And recourse to mathematics is needed in pursuit of
such measurement. Hence the need for mathematics for management. There is saying:
one’s knowledge is one-third if one can only speak on a topic, two-third if on can write in
words and 100% if one can express in numbers. Mathematics the queen of sciences, thus
has a king’s size applications in all walks of life.

1.2 MATHEMATICAL OPERATIONS

We know the basic mathematical operations, namely, addition, subtraction,


multiplication, and division. But many higher order manipulative mechanisms are
available. We have differentiation, integration, programming, simulating, networking,
gaming and other techniques which help in business optimization, i.e., maximizing
revenue/sales/profit or minimizing losses/costs/inputs, subject to constraints and
conditions.

1.3 MANAGEMENT CONCERNS

Management is concerned with income, expense, assets, liabilities, productivity,


efficiency, profitability, costs, volatility or fluctuation in market, market share, media
effectiveness, media reach, brand penetration level, brand equity, effect of pricing
strategy on sales and profitability, motivation levels, motivation and productivity nexus,
coefficient, leadership styles and productivity nexus coefficient, organizational climate
and productivity nexus coefficient, cost of capital, value of firm, capital structure and
value of the firm nexus correlation, credit policy and profitability linkage coefficient,
economic value addition and corporate governance correlation relationship, inventory
levels and cost of inventory nexus, competition levels and business prospects nexus, so

1
on. All these need measurement of the underlying factors to have a greater understanding
of individual factors and their relationship, intense. Management is thus full of
measurement and hence draws heavily from the science of numbers, i.e., mathematics.

1.4 QUANTITATIVE ANALYSIS IN THE PRACTISE OF MANAGEMENT

In the preceding paragraphs a partial list of the concerns of management was


presented. All these concerns require decision making in order to optimize ultimate
results for the organization. Income maximization, expense minimization, asset-quality
maximization, liability cost minimization, productivity maximization, volatility
optimization, market share maximization, optimizing pricing strategy for sales and profit
maximization and so on are called for. To achieve these diverse objectives of the
organization a scientific approach need to be devised.
There are different alternatives to reach the goals set. The right alternative has to
be chosen. Thus a decision-making situation evolves. This situation can be handled
qualitatively as well as quantitatively.
The qualitative approach is heuristic or whimsical or rule-of-thumb based or
intuitive or judgement or trial and error based. The quantitative approach is scientific. It
is open book. It can be tested for efficiency and improved upon.
Quantitative approach is preferred under the following situation:
- the problem is complex, involving many variables
- the interrelationship is complex and multi-faceted
- the problem is repetitive and that quantitative solutions can be repeated several
times thus reducing cost and time

The quantitative approach is preferred for the following reason:


- the language of quantitative approach to decision making is more concise and
precise
- there exists a wealth of mathematical theorems for our use
- forces explicit expression of “ifs & buts”
- enables us to handle any number of variables at a time
- ensured the repeat application to the general n-variable case.

1.5 MODEL OF QUANTITATIVE ANALYSIS

Quantitative analysis is a scientific approach to decision making. As a first step to


decision making, decision model has to be evolved. The decision model depends on two
factors, namely the problem and the problem environment.
1.5.1

The first step in decisions making is defining the problem. The problem (i.e. the
threat, opportunity, etc.) must be fully understood as to its nature, dimensions, intensity
and so on. Take a labour absenteeism. It becomes a problem when it is rampant
affecting work schedules. (one or two absentees, here and there is no problem). What is
its nature? Deliberate absenteeism, absenteeism due to unavoidable causes, absenteeism
as a mark of protest to certain managements attitudes/actions, absenteeism due to

2
healthy/family/social reasons, etc. are indicative of the nature of the problem of
absenteeism. Specific nature must be understood, since each type of absenteeism needs a
different approach to solving the same. Dimensions of absenteeism comes next. How
does it affect production? How does it affect inter-departmental relations, work flow, co-
ordination, morale, etc? What is its impact on turnover and customer satisfaction? What
is the scale of idle capacity resulting from labour absenteeism? Will
motivational/coercive courses work containing absenteeism? There are but a few
dimensions of the problem. How much grave the problem is? Should it be solved
immediately? These indicate the intensity of the problem. The wholesome knowledge of
the problems is very essential so that we can solve the problem. Not all problems can be
known fully. Some keep change from time to time. So, to the extent possible the
decision maker must grasp the problem. This step/function is also called as the
‘intelligence’ function of decision making. All this help to recognize the problem.
Problem definition goes a little for.
In the given situation what is your objective? The decision maker must set the
decision objectives. What constitutes an effective solution? The answer to this question
is what is dealt by decision objectives. Reduce the level of absenteeism from the present
five percent of scheduled work-hours to two percent could be one of the objectives. Let
us deal with deliberate absenteeism and protest absenteeism at the moment, could be
another objective. That is which part of the problem must be solved now is one of the
decision objectives. Let us adopt motivational approaches to solve the deliberate
absenteeism could be another objective. Protest absenteeism must be resolved by hook-
crook approach, otherwise the management may be weakened-this could be another
objective. Objectives give directions to possible courses of actions. And that they lead to
developing alternative solutions that satisfied some broad decisions parameters with
objective formulation, the problem stands defined.
Take another example. There is an opportunity of diversification. If the
opportunity of diversification is seized, backward and forward integration is possible;
competition can be held at bay. But, how to get the necessary resources to take up the
opportunity. Acquisition or “startup from scratch”, internal and external resources, debt
and equity capital etc are the decision issues. In what time frame this must be achieved?
What if competitors try to copy or thwart our efforts? All those make problem complex.
The problem may be a threat. The launch of new designs by competitors is eating
up your market space leading to declining top line and hence bottom line. How to deal
with the problem? Should the firm also launch new models? But that will take some
time. Should it gear up promotion? Should it initiate a price war? How to react to
competitive price war, if that results eventually?

1.5.2 Problem Situation or Environment

Decision situation or environment refers to situational factors that contribute to the


problem, the situational factors that influence the implementation of solution to the
problem and the situational factors that the choice of solution and so on.
The problem may be dynamic or static in nature. Dynamic problems keep changing
from time to time and place to place. Many variables affect the problem and the decision
maker has no control over these. Besides a majority of these variables are external to the

3
organization. Competitive factors, consumer attitudes, technological factors, political
and legal environment etc., are continually changing and constantly influencing business
decisions. Capital expenditure decision, long-run product and marketing decisions, etc.
are typical examples.
On the other hand, some decision situations are static. The variables are known and
can be manipulated by the decision maker. Inventory decisions, employee compensation
decisions, passenger seat-reservation-cum-cancellations etc., are certain examples where
the closed decision system can be used with advantage.
Knowledge about outcome of a course of action is yet another factor constituting
the decision situation. The outcome of certain courses of action can be forecast
accurately, while that of others cannot be. Forecast of return from Government bonds,
leave rentals, etc., can be done with 100% attitude, while that of equity investments
cannot be made accurately.
While this frame work of decision making is common to all decision situations, the
practice of decision making varies from case to case, depending on the decision
environment. Therefore, decision models have to be so developed that they suit the
different decision environment conditions.
The decision environment may be complex, volatile and less amendable to
manipulation or may be simple, static and more amendable to manipulation. Decision
situations with certain environment require closed decision model and with volatile
environment require open decision model. The two decisions models are however, the
extremes of a continuum. Mixed decision models are, therefore, more realistic in nature.
However, an understanding of the open and closed decisions of the models is very
essential in developing a mixed decision system.
A closed decision model has a defined environment and known components. Such
a model can be modeled and its functioning programmed. The model variables are all
known and no random variable enters the model. The model strives to obtain the
optimum outcome. As on approach to business decision making, the closed decision
model has only limited us, since a closed decision model cannot function effectively in
dynamic business environment. However there are certain routine decision situations
where a closed decision model can be used.
An open decision model has on environment that is generally less defined. The
decision maker is not fully aware of the environment he is confronted with. The
variables are too many to corporate in the model. The exact nature of the relationship
among variables is not fully known. Random and unpredictable variables may enter into
the system and effect the outcome. Under such conditions, the decision maker makes
manageable delimitation of the problem. The model goal, variables, alternative courses
of actions, etc., are defined by the maker to the best of his perception and knowledge.
Such limited search is referred to as bounded rationality. The open model has a tendency
to adopt to changing environment and therefore, is increasingly relevant today as a
business needs to adapt on a continuous basis in this dynamic world.

1.6 DECISION MODEL

A decision model is an idealized representation of the problem. Decision


model refers to structured presentation of the problem, solution there to and

4
stimulation of working of the solution. The mode’s purpose is to enable the decision
analyst to forecast the effect of factors crucial to the solution of the problem.

1.6.1 Types of Models

There are different types of models. Ionic and Symbolic models are the prime two
types. Ionic Models are concretized. It is a physical representation of any real life object
on a different scale. Think of a prototype of a plane/car/machine/globe/idol and so on.
Symbolic models re abstract models. A cost curve, a supply curve, marginal revenue
curve, a production possibility curve, etc., is a symbolic model. A forecast profit and lost
account is also a symbolic model. The statement symbolizes summary of financial
effects of commercial activities planned over the next period. An equation is a symbolic
model of variables, their relationship and effect of independent variables on the
dependent variable.
Symbolic models can be classified into: 1) quantitative and qualitative, 2)
standard and customized, 3) probabilistic and deterministic, 4) descriptive and
optimizing, 5) static and dynamic simulative and realistic models.
In a quantitative model all variables are expressed in numbers, while in a qualitative
model some or all of the variables are given only verbal expression.
Standard model is universalized. Computers languages and operating systems are
standardized, while application programmes are customized. Nowadays customized-
standardization is the order, powered by enormous computing powers of computers.
Mass customization is preferred.
Probabilistic model deals with situations where the outcomes of current action are
not known, but their probability distribution is known. Deterministic model deals with
decision situations where certainty of outcomes is taken for granted.
Descriptive model describes the basic relationship between variables. If Rs.10 and
Rs. 12 are the contribution per unit of Product A and Product B, respectively and if a
units of A and b units of B are produced, total contribution is described by: 10a + 12b. In
an optimizing model, ways to maximize the total contribution, given the resources and
consumption and pattern of resources by A and B
Static model assumes, the set of conditions affecting the problem remains
unchanged in a given frame time. Linear programming is a static model. Dynamic
model assumes that the conditions keep changing with time. Dynamic programming is a
dynamic model.
Simulative model tries to replicate the real world situation either through a
computer programme or through Monte Carlo Simulation System or through such other
methods. A realistic model is action play. An advertisement campaign is full scale on.
The firm tries to evaluate the sales profit, awareness and brand equity effectiveness of the
programme.

1.6.2 Model Construction

The construct a model, one needs to know the variables, constants and constraints.
A verbal representation of the relationships among variables is then attempted. Next, a
model is constructed by developing a symbolic statement of the relationships among the

5
variables and then these symbolic relationships are turned into a mathematical model
which is tested with real-world data and evaluated. But all decision situation do not lend
them to modeling.
In the case of decision situations that requires an open system approach, accurate
modeling is utopian, because the decision components are not fully known. Hence, open
decision systems are not perfectly modeled or structured. Depending on individual
situations, models vary from partly structured from, the decision model is an
approximation of the real world situation. Consider the case of sales forecasting, which
is a typical dynamic decision exercise. Sales response constant – ‘r’ advertisement effort
as a function of time – A(t), sales saturation level – ‘M’, sales rate per time – ‘S’, and
sales decay constant – ‘Z’, are some of the relevant variables identified. A simple model
is developed using these variables which runs as follows: ds/dt = r.A(t).(M-S) MZS,
where ds/dt is the rate of change in sales per unit time. Many variables would affect the
rate of sales, but a few only are considered important here. Thus, the above model is only
an approximation.
In certain cases of open decision models, no modeling, however much an
approximation, is possible. In such cases the system remains unstructured. On the
contrary, closed decision models are fully structured for reasons best known. Even here,
constructing a workable model may prove difficult in some cases. The stimulation
technique is adopted then to develop the model.
In case of open decision making models, the environment s risky and stochastic and
in such an environment as outcomes of decisions cannot be known with certainty, the
objective criterion is one of satisfying. In regards to closed systems, on the other hand,
the functioning of the system is highly structured and that optimization is the ultimate
goal.

1.6.3 Constructing a Mathematical Decision Model

Mathematical model is an idealized representation expressed in mathematical


symbols and expressions. A mathematical model of a business problem might be in the
form of a set of equations and related mathematical expressions that describe the essence
of the problem. An economic order quantity model is given by: EOQ = √2AO/C where
A – annual requirement, O – ordering cost and C – carrying cost. A linear programming
model is given by objective function: Say Maximize Z = 10a + 12b, subject to 2a + b < =
60, 3a + 4b < = 120, a, b > = 0, where a and are units of products A and B, respectively to
be produced to maximize total contribution given individual contribution of Rs.10 per
unit of A and Rs.12 per unit of B, the resource constraints being that resource 1 and 2 are
available respectively to the extent of 60 and 120 units only. An internal rate (IRR)
model is given by: -I + Cft/(1 + k)t = 0, where ‘k’ is the IRR to be found, while ‘I’ is the
initial investment, CFt refers to periodic cash flows. To construct a mathematical model
objective of the firm, variables, constants and constraints must be known, besides the
relationship amongst the variables. The relationship may be linear or non-linear. The
EOQ and linear programming models given above are linear relationship based, while the
IRR model is non-linear based as its exponential form.

1.6.4 Variables in Mathematical Models

6
Variables are something whose magnitude can change. These can take different
values. Price, profit, revenue, cost, quantity produced, quantity sold, imports, exports,
income tax, excise duty, national income, savings, consumption, investment, inflation
rate, cost of living, etc., are all variables and can take different values entity to entity,
time to time, place to place and so on.
While variables can take any value, we may give them particular values and
‘freeze’ them in a given context.
Variables may be 1) dependent or independent, 2) stochastic, probabilistic or
deterministic, 3) endogenous or exogenous, 4) continuous or discrete, 5) slack or surplus,
6) choice or random, 7) real or artificial, 8) non-negative or unrestricted, 9) integer or
non-integer variables.
Let E = (P-V)Q –F-D-I, where, E- earnings, P- price per unit, V- variables cost per
unit, Q- quantity produced and sold, F- fixed cost for the period, D- depreciation for the
period and I- interest on borrowed capital. Here, ‘E’ is the dependent variable and all
those on the right side of the equation are independent variables. This is a profit model.
The set of all permissible values that the independent variables can take in a given
context is known as domain of the profit function and the values of E, i.e., profit for the
domain values are called the range of the function.
A stochastic variable takes any value and its value cannot be predicted at all. The
rate of change in sales per time is a stochastic variable. A probabilistic variable takes
values according to a given probability distribution. The number of hit that a web-site
gets per unit time is perhaps Poisson distribution based. A deterministic variable gets a
value that is known beforehand.
Endogenous variables originates from within, say the organization. The ‘Q’, ‘D’ F
and V in the profit function are mostly endogenous, i.e., emanating from inside the
organization. But, P (assumed to be competition triggered) and I (assumed to be money
triggered) are exogenous, emanating from outside environment.
Continuous variables take any value in a continuum, fractional, integer in the
continuum. In normal probability distribution the variable is assumed to be continuous,
while in binomial distribution the variable is discrete, i.e., the variables ‘jumps’, not
‘moves’.
Slack variable is introduced to an inequality of the “<=” type on the left hand side
to convert the same into an equality. A stack thus represents a shortfall being met.
Against this, the magnitude of random variables is the excess of actual over minimum
expected. In respect of a cheaper input, a surplus usage and in respect of a dearer output
a surplus achievement are generally worked out, if possible.
Choice variables, alternatively referred to decision or policy variables are the
variables whose magnitude the organization can pick and choose. Against this, the
magnitude of random variables is not in the hands of the organization. Choice variables
can be also called controllable as to magnitude, while random variables are
uncontrollable.
Real variables are variables that enter the solution set and exist there. Artificial
variables are mathematical artifice and are involved to find a solution, but cannot be part
of the solution.

7
Non-negative variables can take only positive values or zero. They can’t take
negative magnitude. These are referred to a, b>=0. There cannot be situation of -5 units
of product A or -3 units of product B being produced. Unrestricted variables can take
positive and/or negative values.
Integer variables can take only whole number values while non-integer variables
can take any value. Number of chairs/ships produced cannot take fractional value, but
amount of time, consumed can take fractional days/hours.
In constructing a mathematical model all the variables and their nature must be
known.

1.6.5 Constants in Mathematical Models

A constant can take only a specified magnitude. Hence a constant is the antithesis
of a variable. A constant when added to a variable, it is called coefficient of that variable.
However, a coefficient may be symbolic instead of numeric. Suppose let the symbol ‘C’
stand for a given constant and the use of the expression C in lieu of 6C in a model is
perfectly all right and this expression permits greater level of generality. The symbol is
rather a peculiar case. It is suppose to represent a given constant. Yet since it is not
assigned a specific numeric value, it can virtually take any value. In other words, it is a
‘constant’ that is ‘variable’!! Such constants are known as parametric constants or
parameters.
Take the famous equation of Einstein: E=MC2. Here, E- energy produced, M- mass
of the object and C- velocity of light. The velocity of light, C, is a constant. In the
Poisson distribution we use. ‘e’ is a constant, whose value is corrected to four decimals
at 2.7183. The above two constants are classical constants, never change. These are non-
changing constants. But, the parametric constants are constant at the given time and
space. If any of these is changed, the parameters also changes. Thus, a parametric
constant changes with the time and space.
Total cost of production for a given level of output, can be expressed as sum of
fixed cost + Total variable cost. Let the output level be, Q, fixed cost, F, Rs.10, 00, 000
and variable cost, V, Rs.5000 per unit.
Then Total Cost = TC = F + VC OR TC (Q) = 10, 00, 000 + 5000Q
Here, total cost varies as ‘Q’ changes. The fixed cost in total is a constant. But,
these constants are not universal constants. For the time being these are constants. Fixed
cost may change as rent, administrative and depreciation change and unit variable cost
will change if cost of direct material and labour change. Thus these are two constants.
Universal constants and temporal constants.

1.6.6 Enrichment and Evaluation of Model

The model must be enriched and evaluated. Usually, to begin with a mathematical
model is a simple version or an evolutionary design. Step by step the model is enriched
through incorporation of additional relevant elements of (variables and constants) and
conditions and constraints. But too many variables, constants, conditions and constraints
might make model intractable (i.e., incapable of being solved) and imprecise.

8
Enrichment must be attempted to the extent of achieving a valid representation of the
problem.
The validity of the model must be evaluated. Content, context, construct,
concurrent, predictive, convergent, divergent, nomological validities of the model must
be evaluated. Criterion for judging the validity of the model is whether or not the model
measures what it is supposed to measure, predicts the effects of the alternative course of
action with accuracy and so on.
Finally, reliability of the model must be evaluated. A retrospective test may be
performed here. The test involves using historical data to reconstruct the past and then
determining how will the model and resulting solution would have performed if it had
been used. If the results of the model mostly conform with the ex-post results, the model
can be taken to possess reliability. Further, with repeated applications, if the same results
are turned out, the reliability of the model can be taken as established.

1.7 TRADE-OFF

A mathematical model of a business problem aims to achieve the maximum gain


with minimum pain. This is essentially an exercise of trade-off. Risk-Return trade off,
Cost-benefit trade-off, time-cost trade-off and so on are involved.
In Risk-Return trade-off, risk is traded off for return. High risk and high return and
low risk and low return are the order. But a business man wants high returns with low
risk. How is it possible? Impossible the businessman opts for certain level of risk tries to
maximize return for that level of risk or targets a return and attempts to minimize risk as
far as possible for that level of return. This process is called risk-return trade-off.
Similarly, higher benefit involves higher cost and vice versa. Everyone wants high
benefit, but low cost. This is impossible. So, one must decide how much benefit is
wanted and try to reach that level with the least possible cost.
In the case of time-cost trade-off, a project leader wants to complete a project ahead
of normal time. Thus he wants a benefit. But this is accompanied by a cost, cost of
spending up the work. Thus, every gain has a price-tag. The decision maker has to
decide how much cost per unit or utility is optimum. And he settles for that optimum
level of risk-trade, trade-off, cost-benefit trade-off and time-cost trade-off and so on.

Below are given diagrams regarding trade-off.

In the risk-return trade-off, curve AB shows the locus of risk and return. For a unit
rise in return, how much rise in risk takes place? This answer is given by the slope of the
curve. And the slope of the curve is not same throughout the length AB, as AB is not a
straight line, but a curve. But, successive unit rise in return trade-off with increasing
level of risk. The decision maker has to decide the maximum risk in return he is prepared

9
to trade-off. When that level of return is reached, the choice of risk-return combination is
made.
In the time-cost trade-off, initially, for a unit time saved, addition to cost is small.
For successive units of time saved further and further, a higher and further higher
addition to total cost is involved. The decision-maker has to decide how much maximum
additional cost he can afford per unit time saved. Until such level is reached, project
duration is crashed.

Questions
1. Present the application of quantitative analysis in the practice of management.
2. Explain problem definition and problem environment and synthesizing the two.
3. What do you mean by decision model? Explain the types of models and their info
needs.
4. How do you construct a model? How do variables and constant figure in the
model?
5. Explain the concept and types of variables and constants.
6. What is trade-off? How is the same relevant to decision making? How is it
effected?

***

10
LESSON 2
MATHEMATICAL FUNCTIONS AND THEIR APPLICATIONS

A function is called a mapping, or transformation; both words connote the action of


associating one thing with another. A function is a set of ordered pairs with the property
that any particular magnitude of independent variable, say x, uniquely determines the
value of the dependent variables, say y. Thus, a function implies a unique value y for
each x. The converse may or may not be true.

2.1 FUNCTION IN THE FORM OF EQUATION

Function can be presented in the form an equation as below: y = f(x), where y is the
dependent variable and x is the independent variable. [Note, y = f(x), when y is a
function of x, and it does not mean f times x].
x is referred to as the argument of the function and y is called the value of the
function. The set of all permissible values that x can take is known as the domain of the
function. The y value with which an x value is mapped is called the image of that x
value. The set of all images is called the range of the function, which is the set of all
values that y can take.
Let the total cost (C) function of a firm per day is associated with daily output Q;
C= 1500 + 800. The firm has a capacity limit of 100 units of output per day. Then the
domain of the function is the set of the values: 0 = < Q = < 100 or 1500 = < C = < 9500.
By placing the domain on the x-axis and the range on the y-axis, we get the function
in the form of a two-dimensional graph in which the association between x values and y
values is specified by a set of ordered pairs such as, (x1,y1), (x2,y2),…(xn,yn).
C = 1500 when Q = 0 and is 9500 when Q = 100, or 1500 = < C = < 9500.

2.2 DIFFERENT FUNCTIONS AND THEIR GRAPHIC EXPRESSIONS

The expression y = f(x) is a general statement of a function. The actual mapping is


not explicit here.
Constant functions, polynomial functions and relation functions are of one class.

2.2.1 Constant Function takes only one value as its range. y = f(x) = 7; or y = 7; or
f(x) = 7. Regardless the value of x, value of y = 7. Value of y is, perhaps
exogenously determined. Such a function, in the coordinate plane, will appear as
a horizontal straight line. Graph 2.1 gives the constant function.

11
2.2.2 Polynomial Function is a multi-term function. The general form of a single
variable, x, polynomial function is: y = aoxo + a1x1 + a2x2 + a3x3 + …..anxn. Each
form contains a coefficient as well as a non-negative integer power of variables.
[The first two terms can be written as ao + a1x, since, xo = 1 and x1 is commonly
written as x].
Depending on the value of the integer N (which specifies the highest power of x),
several subclasses of polynomial function emerge:
Case of n = 0: y = ao [Constant function]
1
Case of n = 1: y = ao + a1x [Linear function]
Case of n = 2: y = ao + a1x1 + a2x2 [Quadratic function]
1 2 3
Case of n = 3: y = ao + a1x + a2x + a3x [Cubic function]

We can proceed further assigning other values to n. The powers of x are called
exponents. The highest power involved i.e., the value of n, is often called the degree of
the polynomial constant function. The constant function is also called the degenerate or
polynomial of the zero degree polynomial.
Graphs 2.2, 2.3, 2.4, respectively give the linear, quadratic and cubic functions:

2.2.3 Relational functions

A relational function is one which is expressed as a ratio of two polynomials in the


variable x. A function such as

is a polynomial in which y is expressed as a ratio of x – 5 to x2 + 2x + 20

12
(both are polynomials), is a relational function. Any polynomial function is a relational
function, because it can be always expressed as a ratio to 1, which is a constant function.
A special relational function that has quiet an interesting application in business is
the function: y = a/x or xy = a. This function plots as a rectangular hyperbola (Graph
2.5).
Graph 2.5

Since the product of the two terms is always a given constant, this function may be
used to represent average fixed cost curve, a special demand curve where total
expenditure (i.e., price x quantity) is always the same.
The rectangular hyperbola drawn for xy = 1, never meets the axes, for whatever
levels of upward and sideward extensions.

2.2.4 Algebraic and non-algebraic functions

Algebraic and non-algebraic functions are another classification of functions. Any


function expressed in terms of polynomials and/or roots of polynomials is an algebraic
function. The functions so far dealt are all algebraic.
Non-algebraic functions are exponential, logarithmic, trigonometric or so. In an
exponential function, the independent variable appears as the exponent as in the function:
Y = bx . A logarithmic function may be as y = x log b. Non-algebraic functions are also
known as transcendental functions.
Graphs 2.6 and 2.7 give the functions.

13
2.2.5 Linear Functions

Functions can be classified on the basis of number of independent variables. So far


only one independent variable, x, was dealt by us. Instead two, three or any number of
independent variables may be involved. We know production is a function of labour (L)
and (K). so, a production function may be presented as: Q = f(K,L). Consumer utility
can be given as a function of 3 different commodities and the function is y = f(u,v,w).
Functions of more than one variable can be constant, linear or non-linear. Look at
these forms:
Y = a1 + a2 + a3 + a4………an (constant function)
Y = a1x1 + a2x2 + a3x3 + a4x4……..anxn (linear function)

Y = a1x12 + a2x1x2 + a3x22 (quadratic function)


Y = a1x13 + a2x12x2 + a3x1x22 + a4x1x23 (cubic function)

As presented already, linear functions are 1st degree polynomial. These could have
single or more independent variables. A single variable linear function runs like this. Y
= a + bX. A two variable linear function is:
Y = a + bx1 + cx2

2.2.5.1 Concept of slope of a line

Slope refers to the inclination of a line to the x-axis. It gives a measure of the rate
of change in dependent variable, y, for a unit change in independent variable, x. Both the
direction and quantity of change produced are indicated by slope.
Slope of a line can be studied in five ways.
i. Slope is the tangent of the angle made by the line with the x-axis when we move
anti clockwise from the x-axis to the line. If θ is the angle that the line makes with
the positive direction of x-axis, than slope of the = tangent = opposite
side/adjacent side. See the graphs 2.8(1), 2.8(2).

ii. Slope of a line can be measured through ordinates and coordinates. If (x1,y1) and
(x2,y2) are any two points on the line, then slope is given by: (y2-y1)/(x2-x1).

14
iii. When an equation of a line is given as, y = a + bx, then slope is given by ‘b’ in
that equation.
iv. Slope is nothing but the regression coefficient of y on x or byx = ∑xiyi/ ∑xi2 where
xi = xi – x and yi = yi – y.
v. Slope is also studied by amount of change in y, Δy divided by amount of change
in x, Δx. Slope = Δy/Δx.

Illustrations:

1) Suppose three lines make each an angle of 1)30, 2)45 and 60 with the positive
direction of x-axis. Then the slope of the lines are:
1) tan 30° = 1/√3 2) tan 45 = 1; tan 60° = √3
2) Suppose two points on a line are: 2 – 4 and 1, 7. Its slope = (y2 – y1)/(x2 – x1) =
7 – (-4)/1-2 = 11/-1 = -11. The line has negative slope.
3) A machine costs Rs. 1, 00, 000. 5 years after, its value falls to Rs. 60000. If
value is a linear function of time, find the depreciation function.

Solution: Let the value of the machine at year ‘t’ be: v = a + bt. Put v = 1, 00, 000
at t = 0 and v = 60000 at t = 5. We get the following two equations:
i. 1,00,000 = a + b(0) or
a = 1, 00, 000 ….(1)
ii. 60000 = a + b(5) or
a + 5b = 60000 ….(2)

Solving (1) and (2) we get, 5b = -40000 or


B = -8000

The annual rate of depreciation is 8000 and the value of the machine falls
by Rs. 8000 p.a.

2.2.5.2 Linear Cost function and slope thereof

Total cost (TC) = 1500 + 400Q, where ‘Q’ is the quantity produced, and the
constant term is 1500 being the fixed cost and coefficient of Q is 400, which is the slope
of the curve indicating the change in total cost, if output rises or falls by one unit. In the
case of a linear function, the slope is constant at all points on the curve.
Slope is a highly relevant concept in our analysis, slope measures the rate of change
in the dependent variable for a unit change in the independent variable. This is given by
the tangent of the curve at a point.

In graph 2.8 (3) the slope is presented. Slope = ΔY/ΔX

15
We have standardized, Δx as 1 unit. Then Δy, the vertical side of the shaded
triangle, is the slope. And it is constant, for all the four triangles depicted.
Actually slope is the regression coefficient in the regression line. Suppose x
as the quantity and y, the total cost. Let the details be as follows.

Total Maximum
X : 2 3 5 8 12 30 6
Y : 2300 2700 3500 4700 6300 17500 39500
X– : -4 -3 -1 2 6 0-
Y- : -1600 -1200 -400 800 2400 0 -
XY : 6400 3600 400 1600 14400 26400 -
X2 : 16 9 1 4 36 66 -

The regression of X on Y:Y = a + bX


The value of ‘b’ = ∑xy / ∑x2 = 26400/66 = 400. This is the slope of the curve.
a= = 3900 – 400(6) = 1500
So, the regression equation is: Y = 1500 + 400X. Actually this equation is the
same as the TC = 1500 + 400Q, with which we started. Graph 2.9 (1) gives this.
Slope of a curve can also be found by differentiation. We know TC = 1500 +
400Q. Differentiation TC with respect to Q and we will get the slope.
dTC/dQ= 400.
A linear function of the type Y = bX is also possible with no intercept.
Suppose the following pattern of X and Y:
X: 1 2 3 4 5
Y: 100 200 300 400 500

This pattern also depicts a straight line, but its Y intercept is zero. In other
words, the line passes through the origin. Graph 2.9 (2) gives an account of the
same.

16
2.2.5.3 Linear Demand Function and Slope thereof

Suppose the quantity demanded and price ar linearly related and the demand
function is as follows:
Unit price P: 4 5 6 7 8
Quantity Q: 1400 1200 1000 800 600

We are interested in finding the rate of change in Q, for a unit change in P. You
know we have made this rate as 200. We can work out it through the regression
equation, as well as the differentiation route.

Illustration:
First the regression model is attempted.

P: 4 5 6 7 8 30 6

Q: 1400 1200 1000 800 600 5000 1000

P - = p: -2 -1 0 1 2 0 -

Q- = q : 400 200 0 -200 -400 0 -

pq -800 -200 0 -200 -800 -2000 -

p2 4 1 0 1 4 10 -

The slope or repression coefficient = b = ∑pq/∑p2 = -2000/10 = -200.


The constant = Q – bp = 1000 – (-200x6) = 1000 + 1200 = 2200 indicates that,
quantity demanded in indirectly proportional with price, i.e., as P rises Q falls and vice
versa.
Now that we know: Q = 2200 – 200P,
Differentiate Q w. r. t. P, we get:
dQ / dp = -200
This is the same as regression coefficient we got earlier.

Graph 2.10 depicts the demand curve. The curve is downward sloping, indicating
that its slope is negative.

17
2.2.5.4 Linear Supply Function and Slope thereof

Let the supply of goods at different prices be as per the schedule given:
Unit price : P : 1 2 3 4 5
Quantity: Q : 600 800 1000 1200 1400

We can show that the slope is positive and is equal to 200. The equation can be
worked out as before.
P–P=p -2 -1 0 1 2 0
Q –Q = q -400 -200 0 200 400 0
pq 800 200 0 200 800 2000
p2 4 1 0 1 4 10

Let the equation be: Q = a + bp


b = slope = ∑pq / ∑p2 = 2000 / 10 = 200
a = Q – bp = 1000 – 200(3) = 400
The regression equation: Q = 400 + 200p
Graph 2.11 gives the curve
The slope = ΔQ / ΔP = 200

2.2.5.5 Linear Function with plural independent variables

18
2.2.5.6

So far we dealt with linear functions involving only one independent variable. We
can have linear functions with more than one independent variable.
Suppose a firm produces two products A and B. A’s contribution per unit is Rs.
100 and B’s contribution per unit is Rs. 80. Assuming ‘a’ units of A and ‘b’ units B
being produced. The total contribution function is a linear function as follows: 100a +
80b. The contribution function can be expressed in a graph in the form of combinations
of A and B, giving a particular level of contribution. Suppose Rs. 2000 and Rs. 4000
levels of contribution are needed. We can plot in the graph combinations of A and B
giving Rs. 2000 and Rs. 4000 levels of contribution.
Rs. 4000 contribution can be got through 40 units of A or through 50 units of B or
combinations of A and B.

In the graph 2.12, ‘A’ is taken on x-axis and ‘B’ is taken on y-axis. The line joining
40 units of A and 50 units of B, represents combinations of A and B yielding a total
contribution of Rs. 4000. Similarly the line joining 20 units of A and 25 units of B
represents combinations of A and B yielding a total contribution of Rs. 2000. All lines
parallel to these lines represents combination of A and B giving certain levels of total
contribution. For higher total contribution the curve shall be farther from origin and vice
versa.
If 3 independent variables are involved, we can still have graphical version, a 3
dimensional graph. Beyond 3 independent variables, the equation form is only method of
presentation.
Suppose there are ‘n’ independent variables. The general order of linear equation
here is: Y = a0 + a1X1 + a2X2 + a3X3 + ….+ anXn, where X1… Xn are independent
variables and a0, a1, …. an are constants and coefficients.

2.2.6 Non-Linear Functions

19
Non-linear functions are simply the 2nd or further higher degree polynomials. We
know the 2nd degree polynomial is a quadratic function, the 3rd degree polynomial is a
cubic function and so on. You may refer back graphs 2.3 and 2.4 for quadratic and cubic
functions. In graph 2.3 a2 < 0 and the graph appears like a ‘hill’. If a2 > 0, then the
graph will form a ‘Valley’ as shown in graph 2.13.

2.2.6.1 Quadratic Function

Let the demand for a product be given by the quadratic function: Q = P2 – 7P + 10,
where P is price.
We can get the graph plotted by first mapping the function in the form of a table as
below. As P cannot take negative value, we take it as zero first and allow it to rise a unit
at every step and relevant Q values got thus:
P: 0 1 2 3 4 5
Q: 10 4 7 12 19 28

20
Graph 2.14 depicts the function Q = P2 – 7P + 10. The curve does not conform to
normal demand curve, generally speaking. Perhaps this is vanity demand curve. A
quadratic function is different from a quadratic equation. A quadratic equation results,
when a quadratic function is set to zero. That is, the quadratic function, P2 – 7P + 10 if
made as equal to zero, i.e., P2 – 7P + 10 = 0, a quadratic equation results. For the
quadratic equation the roots can be worked out.

2.2.6.2 Cubic Function

Let us continue with the quadratic function of Q = quantity demanded as a function


of P = price, as: Q = P2 – 7P + 10. From this we get the total revenue function, TR = PQ
= P(P2 – 7p + 10) = P3 – 7P2 + 10P. This is a cubic function, a non-linear form of 3rd
order polynomial.
We can plot the curve by assigning values to ‘P’ and getting those of TR. The same
is tabled below:

Price per unit = P = 0 1 2 3 4 5 6 7


Total revenue TR = 0 4 0 -6 -8 0 24 70
(P3 – 7P2 + 10P)

The first derivative of the total revenue function is the marginal revenue function.
We know.

21
That is the rate of change in total revenue for a given instantaneous change in price
2
is 3P – 14P. This is the slope of the curve. It varies from place to place on the curve.

2.2.6.3. Slope in the case of non-linear functions

Slope in the case of non-linear functions is not constant for all points on the curve,
unlike the case with linear functions with linear functions with same slope for all points
on the curve.
Let the cost curve and revenue curve be 500 + 13Q + 2Q2 and 125Q – 2Q2.
The cost curve can be derived as follows:
Q 0 5 10 15 20
TC: 500 615 830 11458 1560
The revenue curve can be derived as follows:
Q: 0 5 10 15 20
TR: 0 575 1050 1425 1700

The slope of the curves are not same at all points. For different magnitude of ‘Q’,
different slope levels exists.
The first derivative of cost function w. r. t. Q gives the slope. It is 13 + 4Q. So,
slope for different ‘Q’ values are as tabled below, obtained by putting the value of Q in
the above slope function:
Q: 0 5 10 15 20
Slope: 13 33 53 73 93
(13 + 4Q)

The first derivative of revenue function w. r. t. Q gives its slope. It is: 125 + 4Q.
So, slope for diff. ‘Q’ values are as tabled below:

Q: 0 5 10 15 20
Slope: 125 105 85 65 45
(125 + 4Q)

22
2.2.6.4 Profit Function

E = Profit = Total Revenue – Total Cost. For the example dealt above, Profit =
125Q – 2Q2 – (500 + 13Q +2Q2) = 125Q – 2Q2 – 500 -13Q -2Q2 = -4Q2 + 112Q – 500.
Profit for 0, 5, 10, 15 & 20 units are: -500, 4, 220, 280 and 140.
The slope of the profit curve i.e., first order derivative with respect to e = dE/dQ of
-4Q2 + 112Q – 500= -8Q + 112. At Q = 0, 5, 10, 15 and 20 the slope is: 112, 72, 32, -8
and -48. Graph 2.18 gives the total profit and marginal profit curves.

What is the profit optimizing output level?

The profit function is: E = -4Q + 112Q – 500.

First derivative dE/dQ =-8Q + 112. Profit maximizing output is given by letting the
first derivative equal to zero and solve for Q. So, -8Q + 112 answer, the 2nd derivative
must be done and if its value is negative, the answer is correct. And d2E/dQ2 = -8. So
our answer is correct.
When marginal profit (i.e., slope of total profit), reaches zero, total profit is
maximum. The quantity at level of profit is profit maximizing quantity. From the graph
2.18 we read it as 14 units, the same as we got earlier algebraically.

23
2.3 ELASTICITY OF DEMAND THROUGH FUNCTIONS

Elasticity is a concept very much used in business. Elasticity of demand, supply,


etc are common parlance terms in business studies.
Elasticity of demand to price of the product, to increase in the income of the
customer, to price of competitor product (known as cross elasticity of demand), tpo
advertisement and promotion campaigns, etc are very useful concepts in businesses. All
these aim to measure the rate of change of quantity demanded to a given change in price,
income, price of competitor product or advertisement and proportional spending.

2.3.1 Price elasticity of demand

Let us now consider price elasticity of demand. Let it be denoted by ‘E’. By


definition E is given by:

= Rate of change in quantity demanded


Rate of change in price

= -ΔQ/Q, Where Q- is the quantity, P – is the price and ΔQ and


ΔP/P ΔP are the changes in Q & P

= -ΔQ x P = P x -ΔQ
Q ΔP Q ΔP

You must remember ΔQ/ΔP is a slope function, slope of quantity w. r. t. price. So,
elasticity of demand is slope of demand curve multiplied by – P/Q. [Note the negative
sign is used as elasticity in both directional and dimensional and as ‘Q’ varies opposite to
the price, the ‘-‘ sign is attenuated to the ΔQ]. For all normal goods and luxury goods the
price elasticity is negative, but conventionally the – ve sign is ignored.
Let the demand function be Q = 280 – 7P. its elasticity is given by –(-P/Q) x
Slope. The slope is simply the first derivative w. r. t. P. So, dQ/dP = -7. The ‘E’ for the
different values of P are computed and given below; and the graph 2.19 gives the demand
curve.

24
When elasticity is > 1, the demand is said to be price elastic and reducing price,
more volume can be sold. When elasticity is < 1, the demand is inelastic. By reducing
price you can not sell more, at the same time by raising the price, your sales volume is
not going to be severely pruned.
When price is reduced from Rs. 30 to Rs. 20, i.e., by 33%, Q i.e., the sales volume
has increased by 100% from 70 to 140. that is, when elasticity was high, a reduction in P
enabled sales. Also, the sales revenue increased from Rs. 2100 to Rs. 2800. When price
elasticity was less than 1 an increase in price will get good revenue. See, when price was
increased from Rs. 10 to Rs. 20, by 100% volume of sales declined only by 33% from
210 to 140. at the same time revenue increased from Rs. 2100 to Rs. 2800. Thus for the
seller, it pays to rise prices when he has price inelastic demand curve. Similarly, when he
forces an elastic demand curve, it pays to reduce price.
The price elasticity is > 1, beyond the price level of Rs. 20. So, any price rise from
Rs. 20 will have the seller with low sales. But price reduction to Rs. 20 from higher price
level will benefit him. The price elasticity is < 1, below the level price of Rs. 20. So, any
price reduction from Rs. 20 will not benefit the seller, but any price rise, movement in
price towards the unit elasticity point from any previous position of elasticity pays off
well the seller.
You can notice the use of slope in computing price elasticity. Price elasticity is
simply slope times P/Q. you can also note, given the slope price elasticity is directly
proportional to price and inversely proportional to quantity.

2.3.2 Income elasticity of demand

Income elasticity of demand process light on the ratio of rate of change in quantity
demand to rate of change in income of the consumers.

Income Elasticity = ΔQ/Q = ΔQ x I = I x ΔQ


Δ I/I Q ΔI Q ΔI

Suppose the following demand income function is present: Q = 500 + 0.05 I, where
Q is quantity demanded and I is income. Then the following table of Q and I ordered
pairs can be worked out.

25
I : 1000 1500 2000 2500
Q : 550 975 600 625
Slope: ΔQ/ΔI : 0.05 0.05 0.05 0.05
Elasticity: (I/Q) x (ΔQ/ΔI): 0.091 0.130 0.167 0.2

[Slope of the Income-demand function is the 1st order derivative of that function.
We know the function is : Q = 500 + 0.05 I. Its 1st order derivative:

2.3.3 Cross elasticity of demand

dQ = 0.05]
dI

The ratio of rate of change in quantity demanded of a product to the rate of change
in price of a related product;
If X and Y are complementary goods (that is use of one needs use of the other too,
like car and car insurance), when the price of Y rises demand for X falls and vice-versa.
Cross Elasticity =ΔQx/Qx = Py x ΔQx
ΔPy/Py Qx ΔPy

If X and Y are competitive goods (that is use of one results in non-use of the other,
like coffee and tea), when the price of Y rises demand for X rises and vice versa.

2.3.4 Advertisement and Promotion Elasticity of Demand

The ratio of rate of change in quantity demanded to rate of change in amount spent
on advertisement and promotion.
Advt. & Promotion Elasticity of demand = (A/Q) (ΔQ/ΔA), where A’ is the amount
spent on advertisement. Here the slope is normally positive and that the demand is elastic
to advertisement and promotion. That is why big companies spend lot on advertisement
promotion.

2.4 MAXIMA AND MINIMA OF A FUNCTION

A function, especially a higher degree function, might be highest for some value of
the independent variable and lowest for some value of the independent variable and there
may be many ‘highs’ and ‘lows’. These ‘highs’ are called ‘maxima’ and the ‘lows’ are
called the ‘minima’. Hence the need to study the maxima and minima of functions.

26
Consider the graph in the interval Q = 1 to 4, f(Q) is increasing to the left of x = 2
and decreasing to the right of Q = 2 and is highest at P, when Q = 2. Consider the
interval, Q = 5 to 7. To the left of Q = 6, f(Q) is decreasing and to the right of O = 6, f(Q)
is increasing and is lowest when Q = 6. When f(Q) is rising, the slope is positive and
when decreasing the slope is negative. And, when slope is zero, either f(x) is maximum
or minimum. So, if the slope is zero, how can we say definitely that f(x) is at its
maximum or at its minimum? Here comes to our rescue the 2nd order derivative or
d2P/dQ2. If the value of d2P/dQ2 < 0, i.e., negative, f(Q) is at its maximum point and if
d2P/dQ2 > 0, i.e., positive, f(Q) is at its minimum point.
Suppose the profit function is, P = Q3/3 – 4Q2 + 12Q.
Find the ‘Q’ for which profit is maximum and the ‘Q’ for which profit is minimum?

Solution

Profit = Q3 – 4Q2 + 12Q


3
2
dP = Q – 8Q + 12 = 0
dQ
i.e., (Q - 2) (Q - 6) = 0
i.e., Q = 2 or Q = 6
of Q2 +8Q + 12 = 2Q – 8
d2P
dQ2

If we put Q = 2, d2p = 2(2) – 8 = 4 -8 = -4. Since, the value < 0, profit is maximum
dQ2

27
at Q = 2. And the profit = 23 – 4(2)2 + 12(2) = 8 – 16 + 24 = 10 2 .
3 3 3
If we put Q = 6, d2P = 2(6) – 8 = 12 – 8 = 4. Since, the value > 0, Profit is minimum
dQ2
at Q = 6. And the profit = 63 – 4(6)2 + 12(6) = 72 – 144 + 72 = 0
3
Questions

1. Give the concept and the types of functions. Illustrate with graphs.
2. What is relational function? Explain the specially of the function y = a/x.
3. What are linear functions? Comment on the slope of linear functions.
4. Explain linear supply and demand functions.
5. What is a non-linear function? Explain the types of the same with
algebraic and graphic illustrations.
6. Comment on the nature of slope in the case of non-linear functions.
7. What is elasticity of demand? Explain measures of computing the same.
8. Explain income and price elasticity of demand.
9. What do you mean by maxima and minima of a function? Explain rules
regarding the same.
10. Explain concepts of domain, image, map, argument, value, dependent
variable of a function.
11. Graph the functions: a) y = -x2 + 5x – 2 and b) y = x2 + 5x – 2 with set of
values -5 < x < 5 as the domain. Which of the functions will have a ‘hill’
and which a ‘valley’. Which factor determine the formation of ‘hill’ or
‘valley’. [Hint: It is the sign of x2 in the quadratic function decide the
formation of ‘hill’ or ‘valley’].
12. A survey shows that there is a linear function between population of a city
and time. In 1986 the population was 68 lakhs, and in 2000 its population
was 75 lakhs. Estimate the population of city in 2005 and in 2010.
(Answer: P = 68 + 0.56T)
13. Find the average rate of change in the supply function, Q = 3p – 4 over the
interval p(1,5). (answer 3)
14. A trader earns Rs. 380 when price = Re 1 per unit, Rs. 660 when price =
Rs. 2 per unit, Rs. 860 when price = Rs. 3 per unit. On plotting the points
(1,380), (2,660) and (3,860) he finds that a quadratic equation may fit the
data. Construct the quadratic equation and estimate his earnings for price
= Rs. 4 per unit. [Hint: construct 3 set of equations find the coefficients
od a, b and c. Earnings at p = 4 = Rs. 980]
15. The total profits of a firm is given by: -2x2 + 400x – 1700, where x =
number of units sold. Find the maximum profit giving sales volume and
profit per unit at that level. (answer: 100 and 183).
16. Cost of production is given by: x2 – 120x + 15000, where x is the units
produced. Find the cost minimizing volume and cost per unit at that level.
17. A firm produces x tons of a product at a total cost given by TC = x3 – 4x2
+ 7x. Find the average cost minimizing output level.
18. Find the maxima and minima of the function: x3/3 – 4x2 + 12x.

28
19. The demand function of a product is: p = 100 – 0.5q2. If price p = 10, find
the elasticity of demand. (answer: 0.05).
20. The demand function is q = 100 – 2p. Find income elasticity of demand
between p = Rs. 25 and Rs. 36.

***

29
LESSON 3
COST-VOLUME-PROFIT RELATIONSHIP

Cost, volume and profit are interrelated. Cost refers to the money value of material,
human and other resources expended in the production of a product or rendering a
service. Cost of production depends on the volume of production. Cost of production is
proportionately, linearly or curvi-linearly related to volume. Hence study pf the pattern
of cost-volume relationship assumes importance. Profit refers to the excess of revenue
over expenses. Profit is a function of volume, cost and price. If the term volume is taken
to mean sales revenue, then profit is influenced by volume and cost. Further as cost itself
is influenced by volume a direct relationship between volume and profit could be
thought. In this lesson varied aspects of the cost-vo9lume profit relationship are
analyzed.

3.1 COST OF OPERATING SYSTEM

There are different costs in the operation of a system. These are: variable cost,
fixed cost, total cost, average cost, average fixed cost, marginal cost, differential cost,
historical cost, future cost, controllable cost, non controllable cost, opportunity cost,
actual cost, out-of-pocket cost, imputed cost, avoidable cost, sunk cost, book cost,
replacement cost, direct cost, indirect cost and so on. These are briefly explained below.
Variable cost varies with the volume of activity in direct proportion. Fixed cost
remains constant irrespective of volume of activity up to certain range. Total cost is the
sum of variable cost and fixed cost. Average cost is the average of total cost or total cost
divided by number of units. Average fixed cost is the total fixed cost divided by number
of units. It forms a rectangular hyperbola. Marginal cost is the change in total cost
caused by a unit change in output. Normally, this is equal to variable cost. Differential
cost is the difference in total cost when different alternative productions methods are
considered. Both change in fixed cost and in variable cost are involved in differential
cost. Historical cost is past cost or cost of events already happened, while future cost is
cost of events to happen. Controllable cost is one which can be modified by executive
actions. Uncontrollable cost is one not lending itself for manipulation. Opportunity cost
is the benefit of next best alternative foregone. Actual cost is cost duly and really
incurred. Out-of-pocket cost is money actually spent while imputed cost is cost attributed
at a job, though not incurred. Avoidable cost is one that can be avoided. Sunk cost is one
already incurred and nothing can be done about the same now. Replacement cost is the
cost involved in replacing a part or segment at current prices, while book cost is cost
historically incurred. Direct cost is cost of direct material, direct labour and direct
expenses incurred in a job, while indirect cost is overhead costs.

30
3.2 PRODUCTION FUNCTION AND COST ANALYSIS

Production function expresses the relationship between input and output. A


production function can be expressed in the form of a table, graph or equation.
Production functions are of several types. One- process production function
meaning only one technically efficient production process, two-process production
function meaning the existence of two efficient processes of production and so on.
Production function in the form of an equation is: Q = f(L,K), where,
Q = units of output (dependent variable),
L = units of labour (independent variable),
K = units of capital (independent variable).

Production function varies with time. There are short run and long run production
functions.

3.2.1 Short-run Production Function

A short run production implies that of the independent variables, a few only can be
changed and the rest can not be changed. So, keeping the unchanged factors of
production at given levels, how the variable factors be changed to optimize the output?
The answer to the question depends on the trend in total product (TP), average product
(AP) and marginal product. As long as additional inputs of variable factor of production
results in addition to TP, i.e., MP > 0, we can go on introducing additional units of that
factor into production.

31
Table 3.1 gives TP, AP and MP of labour with fixed capital
Units of L TP AP MP Units of L TP AP MP
1 10 10.00 10 7 67 9.57 7
2 21 10.50 11 8 72 9.00 5
3 33 11.00 12 9 75 8.33 3
4 43 10.75 10 10 75 7.50 0
5 52 10.40 9 11 73 6.64 -2
6 60 10.00 8 12 70 5.88 -3

From the table it is found that TP is rising till the 9th unit of labour is added, remains
stable between 9th & 10th units of labour and falls when 11th unit of labour is added. That
is, MP is positive till 9th unit, zero at 10th unit and turns negative at 11th unit of labour
onwards. When MP is at its peak, 11 when 3 units of labour are employed, AP is at its
peak too. TP is highest when MP = 0, falls when MP < 0.
As long as MP is rising, it is greater than AP. As MP turns negative, AP starts
falling bit steeply. The first stage of production, i.e., MP > AP, is increasing returns
stage, the second stage, i.e., AP > MP > 0, is decreasing returns stage and the third stage,
i.e., MP < 0, is negative returns stage as far as the factor of production that is being
varied.

32
Graph 3.5, in two periods, gives a vivid picture of the three stages. When MP falls
below AP, inflexion takes place in TP, marking the end of stage I. When MP = 0, TP is
maximum, marking the end of stage II.

3.2.2 Cost of Operation: Relation between cost and activities in the short run

So far we studied the behaviour of output with respect to input in physical units. If
the physical units of inputs are converted into cost, the behaviour of cost and output.
In the short-run, total cost has two components. A fixed component attributed to
factors of production that are kept constant, here capital and a variable component which
is allowed to change, here labour. So, total cost = TC = TFC + TVC, where TFC is total
fixed cost and TVC is total variable cost. Economists consider TVC a non-linear costs
while accountants take it linear cost. The accountants’ version was given earlier in
graphs 3.1 & 3.4. And the economists’ version is given now.
Graph 3.6 gives, in two panels, the short-run cost-output or cost volume
relationship.

33
Marginal Cost (MC) is the rate of change in total cost for a unit change in output.
So, MC is measured through the slope of TC. The slope of TC is different at different
points. Since, TC and TVC are similar in pattern, MC is also measured through the slope
of TVC. This is so, because TFC is constant. Also, using differential calculus, MC can
be studied. This we did in lesson 2.

3.3 LEVEL OF ACTIVITY AND ITS PARAMETERS

Level of activity, i.e., production is influenced by different parameters. The


relationship may be linear, quadratic, etc. These are presented now.
i. Q = a + bL (Linear relationship)
ii. Q = a + bL – cL2 (quadratic relationship)
2 3
iii. Q = a + bL + cL – dL (cubic realationship)
iv. Q = a + bLb (exponential relationship)

3.3.1 Linear Function: Illustration: 1

Suppose production function is linear, for 10 units of labour Q = 45 and for 12 units
of labour Q = 33. Establish relationship between activity and its parameters.
Solution:
45 = a + 10b ------------- (1)
53 = a + 12b ------------- (2)
(1) - (2), -8 = 0 – 2b
or, b = 4.

Putting b = 4, we get a = 5
So, Q = 5 + 4L
MP is the first derivative of Q with respect to L.
So, dQ /dL = 4.

This linear relationship has a limitation. It assumes that unlimited introduction of


the variable factor of production will yield a continuously rising output. It is not so in
reality.

3.3.2 Quadratic Relationship: Illustration: 2

Suppose the activity-parameter relationship is quadratic form given by the equation:


Q = 5 + 4L – L2. Find the MP, output maximizing level of L etc.
Solution
Q = 5 + 4L – L2
MP = dQ = 4 – 2l
dL

To find output maximizing level of L, set MP = 0 and get the value of L. So, 4 – 2L
= 0, or 4 = 2L = 0, or 4 = 2L or L =2. When, L = 2, Q = 5 + 4(2) + 22 = 9.

34
To confirm whether this is the output maximizing level, we have to get d2Q/dL2 and
see whether it is < 0. Here, d2Q/dC2, is less than zero. So, output is maximized at L = 2.

3.3.2 Cubic Relationship: Illustration :3

Let the activity-parameter relationship be cubic. In a particular case it is found to


be: Q = 25 – 18L + 4.5L2 – L3/3. Find the MP, maxima, minima, etc.
Solution:
Q = 25 – 18L + 4.5L2 – L3/3

35
So, MP = dQ/dL = -18 + 9L – L2
By setting dQ/dL = 0, we get, -L2 + 9L + 18 = 0 or,
L2 – 9L – 18 = 0
Or, (L - 6) (L - 3) = 0
Or, L = 6 or 3
If we put L = 6, in the second derivative, we get 9 – 2L = 9 – 12 = -3. Since it is
negative, the function is at one of its maxima when L = 6. At L = 6, total Q = 25 – 18(6)
+ 4.5 (6x6) – (6x6x6)/3 = 25 – 108 + 162 -72 = 7. If we put function L = 3, in the second
derivative, we get, 9 – 2L = 9 – 6 = 3. Since its positive, the function is at one of its
minimum. And at L = 3, Q = 25 – 18(3) + 4.5(3x3) – (3x3x3)/3 = 25 – 54 + 40.5 – 9 =
2.5.
At the maxima, MP should be higher. And it is: 18 + 9(6) – (6x6) = 18 + 54 – 36 =
36. At the minima MP should not be higher than the above figure of 36. Could be less or
equal. And it is: 18 + 9(3) – 3x3 = 18 + 27 – 9 = 36. It is equal here.

3.3.3 Exponential relationship: Illustration : 4

36
Let the activity – parameter relationship be exponential. Say it is: Q = aLb if Q = 4
when L = 4 and Q = 8 when L = 16, find the exponential function.

Solution: a4b = 4 …. (1)


a16b = 8 …. (2)

Taking logarithm on both sides we get,


log a + b log 4 = log 4 …. (3)
log a + b log 16 = log 8 …. (4)

(4) – (3) b log 16 – b log 4 = log 8 – log 4


b (log 16 – log 4) = log 8 – log 4
b = (log 8 – log 4) / (log 16 – log 4)
= (0.9031 – 0.6021) / (1.2041 – 0.6021) = 0.301 / 0.602 = 0.5
Putting the value of b = 0.5 in equation (1), we get
a4(0.5) = 4 or, 2a = 4 or a = 2.
Therefore, the activity-parameter is Q = aL(0.5) = 2L(0.5)

3.4 EMPIRICAL PRODUCTION FUNCTIONS

Certain empirical models of production functions are dealt here.

3.4.1 Cobb – Douglas Production Function

One of the important production functions is given by C. W. Cobb and Douglas.


The function is of the exponential form, with sum of powers of inputs equal to 1. The
function is as follows:
Q = A LbK1-b, (so, b + 1 – b = 1), where,
Q = output, A = a constant, L = Labour, K = Capital and b =

Power a parameter. Actually, the ‘b’ and ‘l – b’, gives the factor-wise shares in total
output.
MP of labour = b(Q/L) and
MP of Capital = l-b(Q/K)

a) Linear Form of Cobb-Douglas Production Function

This is obtained by taking log on both sides of the equation:


Q = ALbK1-b
So, log Q = log A b log L + (l-b) log K

b) Constant Returns to scale of C – D function

37
We know: Q = A LbKl-b. Let K & L are increased by ‘λ’ proportion. Then ‘Q’
increases, say to Q. We can prove Q’ = λ Q, rise in Q is proportionate to rise in
inputs. By original equation,

c) Improved Version of C – D Production Function

The original version of Cobb – Douglas production function is of constant


returns to scale type. Later, this was changed to represent both increasing returns and
decreasing returns to scale as well. The power of L & K were taken as α & β, if α + β
= 1, there is constant return α+ β> 1, there is increasing returns and α + β< 1, there
is decreasing returns to scale. To revised format is therefore, Q = AL α K β.

d) Marginal Product of L and K

3.4.2 Fixed Proportion or Leon tiff Production Function

This production function assumes that the factors of production must be used in
fixed ratio. Excess quantity of either of the input, in relation to other as per the fixed

38
ratio, is useless. The function is: Q = minimum (K/a, L/b), when, K and L refer to capital
and labour, a and b are constants and Q is output. Output depends on smaller of the two
ratios, K/a and L/b.

3.4.3 Linear Programming Production Function

In practice, a firm produces many products and uses many resources, which are
limited by supply, capable of multiple uses and consumed in definite proportion for a
given product. The mix of products must be maximized revenue or profit. In such cases,
linear programming production model is used. (Please refer lesson 4 for LP Production
Function).

3.5 LONG – RUN PRODUCTION FUNCTION

In the long run all factors of production can be altered, enhanced, improved. As a
result, all combinations of inputs giving some level of output by various sets
combinations of inputs.
Graph 3.7 gives isoquant map consisting of different isoquants, each giving a
particular level of output, with different combination of inputs.

100 units of output can be got through different combinations of K and L. L1 and
K1 units of L and K and L2 and K2 units of L and K give some output level. As one
input is decreased, here K from K1 to K2, other input, here L is increased from L1 to L2.

3.5.2 ISO cost Lines


An ISO Cost line is a line depicting several combinations of the inputs involving
same cost. Graph 3.8 gives a set of ISO Cost lines.

39
The line C = 300 indicates several combinations of K & L having same total cost of
Rs.300. Similarly, C = 400 indicates, combinations of K & L having same total cost of
Rs.400.

3.5.3 Cost Minimization given Output

Cost minimization given output is graphically shown in graph 3.9.

First the isoquant of a particular level of output is drawn. Then isocost lines for
different cost levels are drawn. The isocost line that is drawn tangential to the isoquant
gives the minimum cost level for a given output. The isocost = Rs.500 is tangential to
isoquant at point E. OK level of capital and OL level of labour will be used producing
400 units of output at a total cost of Rs.500.

3.5.4 Given Cost, Output Maximization

Maximizing out, given cost is explained below in graph 3.10.

40
Here, given the isoquant line, C = Rs. 500, we should reach the highest possible
isoquant. Draw isoquants. That isoquant, that is farthest from origin, ‘O’, and tangential
to the isocost line gives, the maximum output level given the cost.
Isoquant 400, is farthest from origin and its tangential to isocost line at E. So, given
cost Rs. 500, maximum output is 400 units, produced by OK units of capital and OL units
of labour.

3.6 COST OF OPERATION: RELATION BETWEEN COST AND ACTIVITY


IN THE LONG-RUN

The long-run cost-activity relationship is different from short-run cost activity


relationship. In the long-run there is no fixed cost; so, long-run total cost (LRTC) starts
from origin ‘O’, while short-run total cost (SRTC) starts from the level of fixed cost.
Similarly short-run average cost, long-run average cost, short-run marginal cost and long-
run marginal cost are different.

3.6.1 SRTC AND LRTC

SRTC and LRTC are tangential at point corresponding to output Q1. Except that
output levels, LRTC < SRTC, because in the long run no fixed cost is involved and cost
minimizing opportunities abound.

3.6.2 Short-run average cost (SAC) and Long-run average cost (LAC)

41
LAC curve is tangential to the various SAC curves. LAC curve is flatter ‘U’
shaped, while SAC curves are sharper ‘U’ shaped. LAC < SAC at any level of output.
The minimum point of LAC curve shows the optimum level of activity.

3.6.2 Short-run Marginal Cost (SMC) and Long run Marginal Cost

At each level of activity, a particular SAC is tangent to LAC, the relevance SMC =
LMC. LMC Curve intersects LAC curve at the latter’s minimum point. Some short run
plants have SAC coinciding with the minimum of LAC. At such level of activity SMC =
LMC. So at that point, SAC = LAC = SMC = LMC.

3.7 RELATIONSHIP BETWEEN COST AND LEVEL OF ACTIVITY

The relationship between cost and level of activity may be: 1) linear total cost
relationship, 2) quadratic total cost relationship, 3) cubic total cost relationship or some
other higher polynomial farms. The first three are presented below.

3.7.1 Linear Total Cost (TC) Relationship

42
In linear form, TC = a + b Q, where, TC = total cost, a = a constant or the fixed
cost, b = rate of change in total cost for a unit change in activity or output level. Graph
3.3 gives a vivid graphical version of linear TC function.

Illustration 5
Suppose for 100 units of output the total cost is Rs.1, 02, 000 while for 120 units
the total cost is Rs.1, 18, 000. If linear cost model exists, get the parameters of the
model.

Solution:
Given a + b (100) = Rs.1, 02, 000 … (1)
a + b (120) = Rs.1, 18, 000 … (2)
(2) - (1) we get, 20b = 16000
b = Rs.800
Putting this value in equation 1, we get, a = 22,000
The linear model is: TC = 22000 + 800Q
The average cost is: a/Q + 800. The marginal cost is: b = Rs.800

3.7.2 Quadratic TC relationship

A quadratic equation between total cost and output or activity means TC is


disproportionately changing with output. The rate of rise/fall changes with the output
level
Quadratic TC form may be of two types. (1) TC = a + bQ + cQ2 or (2) TC = a +
bQ - cQ2. The first means an increasingly rising total cost with output. The second type
means a decreasingly rising total cost with output.

In the case of 3.14 model, AC falls at a smaller rate with rising output, while in
3.15, average cost falls at a higher rate with rising output.

43
AC in the case of TC = a + bQ + cQ2, is: a/Q + b + cQ. AC in the case of TC = 1 +
bQ – cQ2 i.e., a/Q + b – cQ. The change in AC in the former case is: -a/Q2 + c and in the
latter is –a/Q2 – C.

Illustration 6:
You are given two total cost functions namely:
1) TC = 10000 + 500Q + 0.3Q2 and
2) TC = 10000 + 500Q – 0.3Q2
Computer, MC, AC, ΔMC and ΔAC at Q = 100.

Solution
i. First case: TC : 10000 + 500Q + 0.3Q2
MC = dTC/ dQ = 500 + 0.6Q
At Q = 100, MC = 500 + 60 = 560
ΔMC = dMC / dQ = 0.6
AC = TC/Q = 10000/Q + 500 + 0.3Q
At Q = 100. AC = 100 + 500 + 30 = 630
ΔAC = dAC/dQ = 10000 / Q2 +3
= -1 + 0.3 = -0.7.

ii. Second case: TC : 10000 + 500Q – 0.3Q2


MC = dTC/dQ = 500 – 0.6Q
At Q = 100, MC = 500 – 60 = 440
ΔMC = dMC/dQ = - 0.6
AC = TC/Q = 10000/Q + 500 -0.3Q
At Q = 100, AC = 100 + 500 – 30 = 570
ΔAC = dAC/dQ = 10000 / Q2 + 3
= -1 – 0.3 = -1.3.

As MC is +ve in first case and negative in second case, AC falls steeper in second
case (at 1.3 per unit) than in first case (at 0.7 per unit).

3.7.2 Cubic TC Relationship


Suppose the cost function is of the cubic type. The general model is: TC = a + bQ –
cQ2 + dQ3. Graph 3.6 gives a model of the cubic total function.

Illustration 7
Let total cost be: TC = 10000 + 500Q – 0.3Q2 + 0.01Q3
Compute TC, AC, MC, AC and MC for output = 100.

Solution:
TC = 10000 + 500x100 – 0.3x100 + 0.01 (100x100x100)
= 10000 + 50000 – 3000 + 10000 = 67000

AC = TC/Q = 10000/Q + 500 – 0.3Q + 0.01Q2

44
= 10000/100 + 500 – 0.3x100 + 0.01 (100x100)
= 100 + 500 – 30 + 100 = 670

AC = dAC/dQ = -10000/Q2 – 0.3 + 0.02Q


= -1 – 0.3 + 2 = 0.7

MC = dTC/dQ = 500 – 0.6Q + 0..03Q2


= 500 – 0.6(100) + 0.03(100x100)
= 500 – 60 + 300 = 740

MC = dMC/dQ = -0.6 + 0.06Q


= -0.06 + 6 = 5.4

We can also find the maximum total cost and minimum total cost output levels.

3.8 COSTS AND PROFIT

The relationship between costs and profits is inverse. As cost rises, profit falls and
vice versa. Profit = Revenue – Cost. We need to bring in revenue function here, to study
relationship between costs and revenues.

Illustration 8
Let ‘p’ be price per unit and ‘x’ be units. A firm’s price function is: P = 300 – 4x
and its cost function is: TC = 500 + 28x. Find profit maximizing output.

Solution:
The total revenue function is: P*(x) = (300-4x) x
= 300x – 4x2
π = Profit = Total Revenue – Total Cost
= (300x – 4x2) – (500 – 28X)
= 4x2 + 272x – 500

Profit maximizing output is compared as follows. First get the 1st


derivative of π w. r. t. x.
dπ / dx = -8x + 272

If we set it to zero, we get, -8x = 272 or x = 34. d2π/dπ2 = -8 and it is –ve.


So, profit is maximized at x = 34.

Alternatively, since, the coefficient of x2 is –ve, the maximum point of x is = (-)


Coefficient of x divided by 2 (Coefficient of x2) l.e., -272/2(-4) = 272/8 = 34. The
general rule of maximum/minimum point is: -b/2a, in the quadratic form, ax2 + bx + c =
0. Whether the function is at maximum or minimum is given by the sign ‘a’. If a > 0,
minimum point and if a < 0, maximum point results.

3.9 RELEVANCE OF MARGINAL COST AND TOTAL COST

45
Here in this section marginal cost, fixed cost, total cost, absorption cost, etc. are
dealt.

Marginal Cost
Marginal cost is the change in aggregate costs when output is changed by a unit.
Total cost is the sum of total fixed cost plus the total variable cost.
The consisting terminology defines marginal cost as ‘the amount at any given
volume of output by which the aggregate costs are changed if this volume of output is
changed by one unit and it is measured by the total of variable cost’. Analyzing this
definition, one can find that with the increase in one unit of output the total cost is
increased and this increase total cost from the existing level to the new level in known as
marginal cost. For example if a company produces 100 units at a total cost of Rs.10, 000
and if the production is increased by 1 more unit, say the total cost is increased to Rs.10,
080. The increase of Rs.80 is the marginal cost at that level of output.
The later part of the definition says that marginal cost is measured by the total of
variable cost. Variable cost is always equal to prime cost plus variable overheads hence
marginal cost is always known as variable cost. If in the above example, if the variable
cost per unit is Rs.80 and fixed cost for the period Rs. 2000, the total cost will be as
follows for 100 units:

Variable cost 100 units @ Rs.80 = 8,000


Fixed cost = 2, 000
Total cost Rs. = 10,000

If the output is increased by one more unit to 101 units the following emerge:
Variable cost for 101 units @ Rs.80 = 8.080
Fixed cost = 2,000
Total cost Rs. = 10,080

The difference in the total cost as well as in total variable cost when the production
is increased by one more unit is Rs. 80 which is the marginal cost.

The institute of Chartered Accountants of England defines marginal as follows:


‘Marginal cost is the very expense (Whether of production, selling or distribution)
incurred by taking of a particular decision’. Thus, it implies that Marginal cost is nothing
but the cost incurred as a result of a particular decision.
Economists define Marginal cost as the additional to total cost or reduction in the
total cost by the production of one more or one unit less.
If we compare the Accountant’s view of marginal cost with that of the economist’s
view we can find out the following differences.
i. Economists take total cost at two different levels of production whereas
the Accountants take the total variable costs only at two different levels.
ii. As viewed by the economists, marginal cost curve would fall in the
beginning and rise rapidly afterwards as the Marginal cost curve will be
always saucer shaped. As per the Accountant’s view, the marginal cost

46
curve would be a horizontal straight line as the Marginal i.e., variable cost
per unit is assumed to be constant.
iii. Economist’s concept of marginal cost is a long-run concept; whereas,
accountant’s view of marginal cost is a short-run concept, because the
total fixed cost and variable cost per unit remaining constant would be
possible only in the short run.

3.9.2 Marginal Costing: Definition

The costing terminology defines marginal costing as “the entertainment of Marginal


costs and of the effect on profit of changes in the volume or type of output by
differentiating between fixed and variable costs”.
The Institute of Cost and Management Accountants of England defines Marginal
costing as follows “Marginal costing is the technique where only the variable costs are
changed to cost units, the fixed cost attributable being written off in full against the
contribution for that period”.
If one analyze the above two definitions, are can find out:
i. Marginal costing is neither a method nor a system, but it it only a
technique i.e., a technique of a decision making.
ii. It involves ascertainment of marginal cost.
iii. Through the ascertainment of marginal cost, we can find out the effect on
profit due to changes in the volume or type of output.
iv. The variable costs are charged to cost units as they only are the product
cost, and
v. The fixed cost, being the period cost should be written off in full against
the contribution for the period. Contribution refers to the difference
between sales and the variable cost of sales. It can be presented as: Sales
– Variable cost = Contribution – Fixed cost = Profit.

3.9.3 Fixed Vs Variable Costs

The whole edifice of marginal costing is built upon one important behaviour of the
cost i.e., variability. On the basis of variability, cost can be classified as (1) Variable (2)
Fixed and (3) Semi-Variable.
i. Variable cost: Variable cost is a cost which in aggregate varies directly
with variations in volume of production. Such cost are supposed to vary
in the same direction and in the same proportion as the variation in the
output. Hence, such costs are also known as ‘product costs’ as additional
production involves only the variable cost. Total variable cost will be
proportionately increasing or decreasing with the volume of production
and variable cost per unit will be constant.
ii. Fixed cost: Fixed cost is that cost which, in aggregate, tends to be
unaffected by variations in the volume of output. The amount of fixed
cost tends to remain constant for all volumes of production within the
installed capacity of the plant. For example, salary to Manager, Rent for
factory etc. shall remain the same even if production goes up or comes

47
down in a period. It is to be remembered that this distinction of variable
or fixed cost is made in the short run only. In the long run all costs are
variable.
iii. Semi – Variable or Semi – fixed cost: This is a cost partly fixed and
partly variable. It has the characteristics of both the fixed and variable
cost. The cost varies with production, but not in the same proportion. E.g.
Maintenance of plant cost, telephone charge, etc., in fact most of the
overhead cost will be in the nature of semi-variable cost only.

3.9.4 Segregation of fixed and variable cost


As the whole marginal costing technique is based upon the differentiation between
variable and fixed cost and as most of the overhead costs are only semi-variable in nature
it becomes necessary to segregate semi-variable expenses into their fixed and variable
elements. The following are the methods employed for this purpose.

a) High and Low Method


This method is based on the analysis of the past records of expenses. The
overheads at the highest level of activity and the lowest level of activity are analyzed
and the variability rate of expenditure is determined as under:

Machine Hours Maintenance Cost


(Rs.)
High 2,000 9,000
Low 1,200 7,800
Difference 800 1,200

Variable expenses per machine hour = Δcost/ ΔHrs = 1200 / 800 = Rs.1.50
Substituting the variable cost in the high or low activity, we can get the fixed cost: High
2,000 Machine hours @ Rs.1.50 = 3,000; Fixed cost Rs.6,000 (i.e., 9000 – 3000).

Low 1,200 Machine hours ~ Rs.1.50 = 1,800; Fixed cost Rs. 6,000 (i.e., 7800 – 1800)

With this information one can project the estimated cost for the future period. For
example, if the estimated machine hours in the future is 3,000 hours, then,
Variable cost 3,000 x 1.50 = Rs.4, 500
Fixed cost = Rs.6, 000
Total cost = Rs.10, 500

This method presumes that the variable portion of the semi-variable cost has linear
relationship i.e. variable cost per unit is assumed to be constant. This may not be true
always.

b) Scatter graph

This method is also known as regression line or the line of best fit. The cost at
all lends are plotted in a graph; for example, the machine hours at different levels or

48
periods are plotted in the ‘X’ axis and the semi-variable cost in the ‘Y’ axis. After
plotting all the points, a straight line is drawn through the points plotted in such a way
that equal distance is maintained as far as possible from these points with almost
equal number of points on each side. The point in the ‘Y’ axis where the straight line
intersects determines the fixed portion of the cost”.

Let the monthly maintenance expenses for 6 months be as follows:


Month Machine Hour Semi Variable cost
(Rs.)
January 4,000 5,600

February 3,000 5,200

March 2,000 4,800

April 6,000 6,400

May 5,000 6,000

June 8,000 7,200

The least square method uses a mathematical approach to find out the variable
and fixed expenses. The method of least sequences is based on the regression

49
equality y = a + bx where ‘a’ is the fixed element and ‘b’ is the degree of variability
and y the total cost. Applying this basic equation and given a set of observations, n,
two simultaneous equations can be developed that will fit a regression line to a linear
array of data. The equations are

∑Y = Na + bX∑ …. (1)
∑XY = a∑X + bX2 …. (2)

Where a = fixed cost


b = variable rate
n = number of observations
X = activity measure (hours etc)
Y = total mixed cost observed
Solving the equations for a & b will get us to find the fixed cost and variable
cost.

c) Least Square Computation


Let x be the hrs and y be the amount. The figures are taken here is Rs. 100
omitting the two zeros.

S.No X Y XY X2

1 40 56 2240 1600

2 30 52 1560 900

3 20 48 960 400

4 60 64 3840 3600

5 50 60 3000 2500

6 80 72 5760 6400
Total 280 352 17360 15400

N=6
∑X = 280
∑Y = 352
∑XY = 17360
∑X2 = 15400

Putting these values in the equations, we get.


6a + 280b = 352 …. (1)
280a + 15400b = 17360 …. (2)
(1) x 14 84a + 3920b = 4928 …. (3)
(2) x 0.3 84a + 4620b = 5208 …. (4)

50
(3) – (3) 700b = 280
b = 0.4

Putting this value in equation (1), we get:


6a + 280(.4) = 352
or 6a = 352 – 112
6a = 240
a = 40
Y intercept = 40 hundred or Rs.4000
The equation is: Y = 4000 + 0.4X

3.9.5 Marginal cost vs. Absorption cost

Marginal costing has to be differentiated from absorption costing which is also


known as total costing or full costing. Absorption costing is ‘the practice of charging all
costs to operations, products or processes’. Under this approach, the total cost is the sum
total of all costs, variable or fixed, the total cost per unit will remain constant only when
the level of output or mixture remains the same from period to period. As these factors
are varying, the cost also would be varying from period to period and hence creates
difficulties in making decisions regarding price fixing, activity level planning, profit
estimating etc. There are certain specific difficulties in absorption costing as discussed
below.

1) It involves apportionment of fixed overheads between products, processes etc. to


find at the total cost and apportionment can be done in any of the various methods
available and as such an element of arbitrariness is introduced in cost
determination resulting in over or under-absorption of costs.
2) In absorption costing as the closing stock of finished goods and work-in-progress
are valued at total cost which includes variables and fixed cost, the cost of one
period is carried forward to the other period, through the value of closing stock.
Fixed cost, strictly speaking is a period cost and should be only recovered from
the revenue of the same period in which it is incurred.
3) Carrying forward of the cost of one period to another as above distorts the results
of both the years, vitiating the very purpose of costing and financial reporting.
4) As a result, cost ascertainment and profit planning become difficult under
absorption costing and this necessitates the application of marginal costing for
managerial decision-making. Hence the significance of marginal costing.

3.10 RELEVANT COSTS FOR DECISION MAKING

There are different cost concepts but all are not relevant in all decision situations.
Table I gives pairs of relevant and irrelevant costs for decision making.
TABLE I: Relevant and Irrelevant Costs

Relevant Costs Irrelevant Costs

51
Future cost Historical cost
Variable cost Fixed cost
Differential cost Total cost
Controllable cost Non-controllable cost
Opportunity cost Actual cost
Out-of-pocket cost Imputed cost
Avoidable cost Unavoidable cost
Replacement cost Book cost

Relevant cost is one that is significant in a given decision making situation. It


influences the choice of the decision maker. All other costs which are insignificant in the
given contest are irrelevant costs. For instance, a concern and interested in achieving cost
efficiency would have to initiate actions to control and contain costs that are to be
incurred. On the other hand, costs that were already incurred cannot be controlled or
contained now. In other words, future costs are relevant for decision making while past
costs are not. In a similar way, variable cost, differential cost, opportunity cost, etc. are
not.
Managerial decisions are varied, complex and conflicting in nature. This requires
the manager to adopt a wholesome approach in a given situation. That is, as the decisions
are varied various cost concepts have to be used in different circumstances. Again, as the
decisions are complex, more than one cost concept may have to be depended on in one
particular situation. Further, as some decisions are mutually conflicting, non-cost
considerations, including the so-called irrelevant cost-concepts, have to be taken into
account as well. In other words, the above cost classification is not absolute, but only
relative. In other words, the above cost classification is not absolute, but only relative.
Thus a particular cost concept or concepts would be relevant in one situation and
irrelevant in another. But the above classification mostly holds goods.

3.10.1 Future vs. Historical Cost

Planning and control are two important managerial functions. Both are forward
looking as management decisions relate to the future. One can plan for the future only.
Similarly, one can control what is to happen rather than what has already happened. It
follows that it is the future cost and not the historical cost that is relevant for decision –
making.
Future cost is associated with future events and course of actions. It includes,
among others, standard cost, budgeted cost and costs to be incurred. Standard cost serve
as yardsticks against which the incurred cost is compared to assess performance. The
deviation between the standard and actual is analyzed for its cause and effect control.
Thus, future cost is relevant to management decision making.
Historical cost, on the other hand, cannot be influenced or controlled since it relates
to the past. However, it should not be concluded that historical cost is totally irrelevant.
In fact, it serves as a guide to planning, it helps constitutes the planning to some extent.

3.10.2 Variable vs. Fixed Cost

52
Level of activity and profit planning, price determination, product mix planning,
make or buy decisions, sell or process further, etc, are some of the important managerial
decision areas. In these decisions, variable cost plays a key role and hence it is a relevant
cost. Fixed cost is irrelevant in this context.
Variable cost changes with the level and mode of operation. It may vary in linear
or non-linear proportion. The management has to decide the particular level of operation
or mix of products where there is cost advantage. In quoting special prices to special
consumers and in fixing minimum prices for products, variable cost is an essential factor.
Thus it is relevant to managerial decision making. On the contrary, fixed cost, for the
total cost is not affected by fixed cost. However, in the long run all costs are variable and
long run decisions are taken without highlighting the variability of costs with volume.

3.103 Differential vs. Total Cost

To decide is to choose from between or from among alternatives. This involves


evaluation of each alternative against the other. But management decision is influenced
by the difference in the costs of the alternatives rather than by the total costs of the
alternatives. This means the differential cost is relevant for decision making, while the
total cost is irrelevant. When we consider the differential cost the distinction between
variable cost and fixed cost is not entertained. That is the differential variable cost and
differential fixed cost together influence decision making.

3.10.4 Controllable vs. Non-controllable Cost

Controllable cost refers to a cost whose incurrence and its scale and time are
subject to executive authority. Non-controllable cost is thrust on the management by the
external factors. Litigation costs, cost of meeting statutory requirements etc. are
examples of non-controllable cost. Non-controllable cost is irrelevant decision making as
it can be manipulated by executive decisions. Such costs can be planned and also
controlled.

3.10.5 Opportunity vs. Actual Cost

Opportunity cost refers to the value of benefits of the next best foregone
alternative. Simply, it means the income from the next best missed alternative. It follows
that opportunity cost would arise only when there are alternatives and decision making
itself arises only when there are alternatives. Hence in all decision situations the
opportunity cost concept emanates. It is specially used in resource allocation problems,
when the resource in question.

3.10.6 Out-of-pocket vs. Imputed Cost

Out-of-pocket cost refers to the cost that involves an actual outlay of cash
immediately or in the near future. Material cost, labour cost and interest on borrowed
funds are examples of out of pocket cost. On the other hand, imputed costs does not
involve actual outlay. It refers to assumed costs of using owned resources for the

53
business carried on by the owner himself. Assumed interest on owner’s capital, assumed
remuneration for owner’s service, assumed rental changes on the owner’s building used
for the business, etc. are some typical examples. For decision making the imputed cost is
irrelevant as it does not affect the business funds and the cost of operation of his business
is not altered by the imputed cost. On the other hand, cost involving actual outlay is
relevant as the manager can plan it and exercise control over it.

3.10.7 A voidable vs. Sunk Cost

If a manager decides to raise additional finance by issuing equity shares instead of


debentures, he can avoid payment of interest. Even if the debentures are already issued
the manager can raise additional equity capital and redeem the debentures, thereby
avoiding interest. Avoidable cost can be of two types – that which can be avoided once
and for all and that which can be avoided for the time being. Both types are relevant to
decision making. It is a future cost.
Sunk cost results from past decisions of actions. It is already incurred and that it is
a type of historical cost. Cost of production and storage of stale or unrealizable stock,
investment in fixed assets are typical examples of sunk cost. But it is a past cost and
future decision would not be influenced by it. Thus sunk cost is irrelevant to decision
making, being unavoidable and unalterable.

3.10.8 Replacement vs. Book Cost

Replacement cost is a hypothetical cost but is relevant to decision making. It refers


to the cost of replacing an asset by a new one. It is obviously a future cost. Assets
replacements often taken place in businesses due to technological changes, wear and tear
and other factors. In this areas of decision making, the replacement cost of existing
assets is an important factor. In the factor market, different brands and varieties of assets
would be available, each involving a different cost. Hence, with the right decision the
management can control the scale and time of incurrence of this cost which is relevant to
decision making.
On the other hand, book cost – as recorded in the accounting records – is irrelevant
in replacement decisions. It is useful only for tax and financial accounting. From the
decision making point of view it is of no significance, as it is only a past cost. Further,
whether the book cost is the original cost or the ‘written down cost’ makes no difference.
To manager deciding to replace a particular asset its book cost, whether adjusted or
unadjusted for depreciation, is irrelevant. It is the replacement cost that is relevant
because it is a future cost.
Thus, with the availability of a number of cost concepts, the manager has to first
ascertain which costs are relevant and which irrelevant in a given situation. It can be
observed from the analysis that all relevant costs have one common feature – they relate
to the ‘future’. It can be stated that future costs are relevant costs. But this has to be
qualified, as all future costs cannot be planned and controlled – for instance, the non-
controllable and unavoidable costs which are beyond the manager’s authority,
comprehension and hence control. One can conclude that all controllable future costs are
relevant costs and all other costs are irrelevant for decision making.

54
3.11 BREAK-EVEN ANALYSIS AND ITS USES

The relationship among the volume of business, cost and profit is very significant.
The profit depends upon volume of production and the type of production. Volume is the
single largest factor which influences cost and profit. Break-even analysis goes deep into
analyzing the relationship among cost, volume and profit. Hence it is also called C-V-P
analysis.
The objective of break-even analysis (otherwise called cost-volume-profit analysis)
is to assist profit planning, cost control and decision making. With a view to maximize
profit, the management has to expand output which will affect the variable cost,
contribution and profit. Hence the management has to know the effect of an increase in
output on its profit. This information will help the management in its budgeting and
budgetary control. Decisions relating to the volume of production, type of production
and pricing are influenced by knowledge of the cost behavior and hence the cost-volume-
profit relationship becomes vital for decision-making.

3.11.1 Break-even Point

In the C-V-P analysis, the determination of break-even point is an important step


and break-even analysis becomes the key component of C-V-P analysis. The break-even
point is that level of production and sales at which there is neither a profit nor a loss. In
other words this is the level of production and sales at which the sales value equals to its
total cost. At this point of sales, contribution is equal to fixed cost. Any sales above the
break even sales only will yield profit and if the actual sales is less than the break-even
sales, it would result in a loss because of the fact that the contribution is not sufficient to
cover the fixed cost. That portion of fixed cost could not be covered is the loss.

The break-even point of sales can be computed by any one of the formula given
below:

Break-Even point (in Rs.) = Fixed cost x Sales = FxS


Sales – Variable cost S–V

or = Fixed cost x Sales = FxX


Contribution C

or = Fixed cost x Sales = FxS


Fixed cost + profit F+P

or = Fixed cost = F or F
1 Variable cost 1-V/S P/V ratio
Sales
Break-even point (in units)= Total fixed cost/ Contribution per unit

55
Illustration 9
Calculate Break-even sales from the following figures.
Sales Rs. 1,50,000
Fixed cost Rs. 37,500
Direct Material Rs. 50,000
Direct Labour Rs. 30,000
Direct Expenses Rs. 20,000

Solution
Break-even sales = F*S / S – V
= 37,500 X 1, 50, 000 = Rs. 1,12,500
1, 50, 000 – 1,00,000
Verification
B.E Sales = 1, 12, 500
Less V.C at that level 75,000
Contribution 37,500
Less: Fixed cost 37,500
Profit NIL
3.11.2 Profit Volume Ratio (Contribution – Sales Ratios)

The profit volume ratio (P/V ratio) is the ratio of contribution to sales and indicates
the relative profitability of different processes, departments etc. It reveals the effect on
profit of charges in the volume without an increase in the fixed cost would yield a higher
profit and vice-versa. It is computed as follows:

i. P/V Ratio = Contribution x 100 = C x 100


Sales S
If selling price and variable cost are constant, then
P/V Ratio = Change in Contribution x 100 = ΔC x 100
Change in Sales ΔS

If the selling price, variable cost and fixed cost are constant, then
P/V Ratio = Change in Profit x 100 = ΔP x 100
Change in Sales ΔS

Using the same relationship it can be derived that:


ii. Sales = Contribution / P/V Ratio
iii. Contribution = Sales x P/V ratio
iv. Sales required to make a desire profit = Fixed cost + Desired profit
P/V ratio
v. Sales at Break-even profit = Fixed cost
P/V ratio

3.11.3 Properties of P/V Ratio:

i) It is the result of linking contribution to sales.

56
ii) It remains constant so long as selling price and variable cost per unit remain
constant or fluctuate proportionately.
iii) It is unaffected by any change in level of activity.
iv) It is unaffected by any fluctuation in fixed cost.

3.11.4 Uses of P/V ratio

i) It helps in determining the Break-even point. BEP = FC/P/V ratio


ii) It helps in determining contribution / profit at any volume of sales
iii) (Contribution = S x P/V ratio; profit = C – F)
iv) It helps in determining the sales volume to earn the desired profit required.
v) Sales = (F + P) / P/V, Where P is desired profit
vi) Helps to determine the profitability of products/ product mix.

Any improvement in P/V ratio would result in increased profit and profitability.
To improve the P/V ratio, either you have to reduce marginal cost, increase the selling
price or by selling more profitable product or product mix.

Illustration 10: A factory produces 600 units per month. Selling price is Rs.240 per
unit; variable cost is Rs.160 per unit. Fixed cost per month is Rs.16, 000.
i) Calculate the estimated profit in a month wherein 480 units are produced and
ii) Find out the sales to make a profit of Rs.14, 000 per month.

Solution:
Selling price per unit = Rs.240
Variable cost per unit = Rs.160
Contribution unit = Rs. 80

P/V ratio = C/S x 100 = 80 / 240n x 100 = 331/3%


iii) Profit on 480 units: Rs.
Sales (480 x 240) = 1, 15, 200
Contribution 331/3% on sales = 38,400
Profit = C – F = 38400 – 16000 = 22400

iv) Sales to make a profit of Rs.14,000


Required profit = 14,000
Add: Fixed cost = 16,000
Required contribution = 30,000

Sales required = Required Contribution = 30000


P/V ratio 331/3%
= Rs.90, 000

3.11.5 Margin of Safety

57
The excess of sales over break-even sales is margin of safety. In other words, it is
the excess of actual sales over break-even sales. Margin of safety is calculated using a
formula as shown below.

M.S = Profit / P/V

Margin of safety ratio = M.S.S x 100


Actual Sales

Or

Actual Sales – BES X 100


Actual Sales

Properties
i) Margin of safety indicates the strength of business.
ii) High margin indicates that concern would make profits even if there should be a
fall in production or sales.
iii) Low margin of safety indicates high fixed cost and profits cannot be made unless
volume of sales is increased or selling price is raised or costs reduced or a
more profitable product is substituted.

Illustration 11
M.Ltd. manufactures one uniform product X. The following figures are available
for two successive years.
Year I Year II
Rs. Rs.

Sales 1, 00, 000 1, 20, 000


Fixed cost 30,000 40,000
Variable cost 50,000 65,000
The directors want information about the BEP, P/V ratio and margin of safety.

Solution:
I Year: Sales Rs. 1, 00, 000
Less: Variable cost Rs. 50,000
Contribution 50,000
Fixed cost 30,000
Profit 20,000

B.E Sales = F x S = 30,000 x 1, 00, 000 = Rs.60, 000


C 50, 000

P/V ratio = C x 100 = 50, 000 x 100 = 50%


S 1, 00, 000

58
Margin of safety = Actual sales – BE sales
= 1, 00, 000 – 60, 000 = Rs. 40, 000

Margin of Safety Ratio = Margin of Safety sales x 100


Actual sales

= 40, 000 x 100 = 40%


1, 00, 000

Alternatively M.S = Profit = 20, 000 x 100 = 40, 000


P/V ratio 50

II Year: Sales Rs. 1, 20, 000


Less: Variable cost Rs. 65, 000
Contribution 55, 000
Fixed cost 40, 000
Profit 15, 000

B.E. Sales = F x S = 40, 000 x 1, 20, 000 = Rs.87, 273


C 55, 000

M.S = 1, 20, 000 – 87, 273 = Rs.32, 727.

M.S ratio = 32,727 x 100 = 27.27%


1, 20, 000

Illustration 12
MV Ltd. a multi product company furnishes you the following data relating to the
year 2001.
First half of the year Second half of the year
Rs. Rs.
Sales 45, 000 50, 000
Total cost 40, 000 43, 000

Assuming that there is no change in prices and variable costs and that the fixed
expenses are incurred equally in the two half-year periods. Calculate for the year 2001: I)
the P/V ratio, ii) fixed expenses, iii) break-even sales and iv) percentage of margin of
safety.

Solution
i) P/V ratio
Sales Total Cost Profit
Rs. Rs. Rs.
Second half year 50, 000 43, 000 7, 000
First half year 45, 000 40, 000 5, 000
Differences 5, 000 3, 000 2, 000

59
P/V ratio = Difference in Profit x 100
Difference in Sales
= 2, 000 x 100 = 40%
5, 000

ii) Fixed expenses


Sales for the I half year 45, 000

Contribution 1 half year 45, 000 x 40 = 18, 000

100
Less: profit for the I year 5, 000
Fixed expenses 13, 000

Therefore fixed expenses for 1 year: 13, 0000 x 2 = Rs.26, 000

iii) BE Sales

BE Sales = FC = 26, 000 = 26, 000 x 100


P/V ratio 40/ 100 40
= Rs.65, 000

iv) Margin of safety percentage


MS = AS – BES = 95, 000 – 65, 000 = Rs. 30, 000
% of MS = (30,000 / 95, 000) X 100 = 31.6%

Illustration 13
Two business units X Ltd. and Y Ltd. sell the same type of product in the same
type of market. Their budgeted profit and loss accounts for the coming year are as
follows:
X Ltd. Y Ltd.
Rs. Rs.
Sales 1, 50, 000 1, 50, 000
Less: Variable costs 1, 20, 000 1, 00, 000
Fixed cost 15, 000 1, 35, 000 35, 000 1, 35, 000
15, 000 15, 000

You are required to:

a) Calculate the break-even point of each business.


b) Calculate the sales volume at which the business will earn Rs.5000/- profit.
c) State which business is likely to earn greater profits in conditions of
i) Heavy demand for the product
ii) Low demand for the product

60
Solution
a) i) P/V ratio = FC + Profit x 100
Sales
For X Ltd. = 15, 000 + 15, 000 x 100 = 20%
1, 50, 000

For Y Ltd. = 35, 000 + 15, 000 x 100 = 331/3%


1, 50, 000
ii) Break-even sales = FC
PV Ratio

For X Ltd. = 15, 000 = 15, 000 x 100 = Rs.75, 000.


20/100 20

For Y Ltd. = 35, 000 = 35, 000 x 300 = Rs. 1, 05, 000
100/300 100

iii) Sales required to make a profit of Rs.5, 000

For X Ltd. = FC + Profit = 15, 000 + 5, 000


P/V 20/100

= 20, 000 X 100 / 2 = Rs.1, 00, 000


For Y Ltd. = FC + Profit = 35, 000 + 5, 000
P/V 100/300
= 40,000 X 300/ 100 = Rs.1, 20, 000

v) In conditions of heavy demand a concern with larger P/V ratio can earn greater
profit because of greater contribution. Thus Y Ltd. is likely to earn greater
profit.
vi) In conditions of low demand, a concern with lower break-even point is likely to
earn more profit because it will start earning profit from a lower level of sales.
X Ltd. will start earning its profit from Rs.75, 000 sales level whereas Y Ltd
will start earning profit only from Rs.1, 05, 000 sales level.

3.11.6 Break-even Chart

The relationship between costs, sales and profits can be shown in the form of a
chart. Such a chart not only depicts the level of activity, where there will be neither loss
nor profit but also shows the profit or loss at various levels of activity.
On the X axis of the graph is plotted the volume of production or the quantities of
sales and on the Y axis costs and sales revenues are plotted. The fixed cost line is drawn
parallel to the X-axis and the variable cost line is depicted above the fixed cost line which
shows that the cost is increasing with the increase in the volume of output. The line can
also be regarded as the total cost line because it starts from the point where the variable

61
cost is zero and certain fixed cost has been incurred. Thereafter the sales revenue are
plotted from the origin and a line is drawn up which goes in the upward direction with the
increase in production/ sales. The two lines – total cost line and total sales line shall
intersect each other at one point and a perpendicular can be drawn from this point to find
out the level of output where the business is at no profit no loss position, called break
even sales. Sales below this point will result in loss and above this point will produce
profit.

Illustration 14:
From the following data compare the break-even point by a break-even chart.

Selling price per unit Rs.10


Variable cost per unit Rs.6
Total fixed cost Rs.2000

Solution

Take cost and revenue on Y axis and output on X axis. At Rs.2000 land draw the
horizontal line for fixed cost. The BE point = FC/ contribution per unit = 2000/4 = 500
units. To draw total revenue line we can take origin as one point and another at 200
units, yielding a revenue of Rs.2000.
To draw the TC line we start at fixed cost point at Rs.2000, which os total cost for
zero output. At 200 units the TC is 2000 + 200(6) = 3200. So the two points are (0,
2000) and (200, 3200). The line is drawn . TR and TC intersect at ‘Q’ which gives BE
position. A perpendicular to x-axis gives BE production (500 units) and to y-axis gives
BE sales (Rs.5000).

62
3.11.7 Assumptions of Break-Even Charts

i) Fixed costs remain constant at all levels and they do not vary with the production.
ii) Variable cost varies in the same proportion and in the same direction.
iii) All costs are capable of being bifurcated into fixed and variable elements.
iv) Selling price remains constant at all levels of production and sales.
v) Cost and revenue depend only on volume and not on any other factor.
vi) Production and sales figures are identical.
vii) Either only one product is produced or the product mix is constant.

3.11.8 Advantages of Break-Even Charts

i) Provides detailed and clearly understandable information regarding the break-


even point, margin of safety, profit at different sales level, etc.
ii) Profitability of the product can be known.
iii) Effect of changes in cost and selling price can be demonstrated.
iv) Through budget break-even chart, planning and control of cost is possible.
v) Economy and efficiency in operation can be achieved.

3.11.9 Limitations of Break-Even Charts

i) Break even charts are based on certain assumptions as shown above but they are
not valid in all circumstances. Fixed costs and variable costs cannot be
correctly segregated, fixed costs do not always remain constant; variable cost
do not always vary proportionately; production and sales are always not equal
and so on. Hence, the concept is based on unrealistic assumptions.
ii) Break even charts provide only limited information. If we have to study the effect
of changes of fixed cost, variable cost and selling prices a number of charts
will have to be drawn up.
iii) There is no necessity for a break-even chart because (a) simple tabulation of data
is sufficient (b) conclusive decisions are not possible (c) difficult for a layman
to understand and (d) there is no basis for comparative efficiency of different
units.

In spite of all these limitations, it is a useful tool for the management for analyzing
its different day to day problems in decision making.

Questions

1) Explain different costs of operating system.


2) Explain production function and behavior of TP, MP, AP with an example.
3) Graph the behavior of TP, MP and AP.
4) Graph the behavior of TC, TTC, TVC, AC, AFC and AVC.

63
5) With 15 units of labour 45 units of output and obtained while 20 units of labour
yield 55 units of output. Find the marginal productivity of x and output
optimizing level of x.
6) A quadratic production function is as follows: Q = -20x2 + 160x + 400, where Q =
output and x = an input. Find the marginal productivity of x and output
optimizing level of x.
7) A cubic production function present is as follows: Q = 20x + 4x2 – x3/3, where Q
= output and x = an input. Find units of x that gives highest Q and maximum
value of marginal productivity of x.
8) A production function is as follows: Q = 2HLK – AL2 – BK2. Find the MP of L
and MP of K and from that marginal rate of substitution of L for K at any point on
the iso-product curve.
9) Explain the different forms of production function.
10) Explain the features of Cobb-Douglas production function.
11) Explain the difference in behaviour of SRTC and LRTC.
12) Explain the derivation of LAC, LMC and LTC.
13) Total cost function is: TC = x2 – 120x + 15000. Find the MC and cost minimizing
total output.
14) Comment on the behaviour of cost of the two function: C1 = -x2 + 60x + 6000 and
C2 = x2 – 60x + 6000.
15) The profit function is as follows: P, and x = units of input. Find the profit
maximizing output.
16) The cost function for a firm producing x units is: TC = 5x + 350 and its revenue
function is: TR = 5x – x2. Find the demand function, break-even–point, profit
maximizing output and sales maximizing output. [Hint: i) For break-even point,
solve for x, by equating TC = TR].
17) A review, made by the top management of Sweat and Struggle Ltd. which makes
only one product, of the result of the first quarter of the year revealed the
following:
Sales in unit 10, 000
Loss in Rs. 10, 000
Fixed cost (for the year Rs. 1, 20, 000) in Rs.30, 000
Variable cost per unit in Rs. 8

The finance manager who feels perturbed suggests, that the company should
atleast break even in the second quarter with a drive for increases sales. Towards
this, the company should introduce a better packing which will increase the cost
by Rs. 0.50 per unit.

The sales manager has an alternate proposal. For the second quarter additional
sales promotion expenses can be increased to the extent of Rs.5, 000 and a profit
of Rs.5, 000 can be aimed at the period with increased sales.

The production manager feels otherwise. To improve the demand, the selling
price per unit has to be reduced by 3%. As a result the sales volume can be
increased to attain a profit level of Rs.4, 000 for the quarter.

64
The managing director asks you as a Cost Accountant to evaluate these three
proposals and calculate the additional sales volume that would be required in each
case, in order to help him take a decision.

18) Details about the single product marketed by a company are as under:
Per unit Rs.
Selling price 100
Directory material 60
Direct labour 10
Variable overheads 10
No. of units sold in the year 5035. Pursuant to an agreement reached with the Employees
union there would be next year a 10% increase in wages, across the board for all those
directly engaged in production.
Work out:
i) How many more units have to be sold next year to maintain the same quantum of
profit?
ii) In else, by what percentage the selling price has to be raised to maintain the same
P/V ratio?
19) Ram Dass Pvt. Ltd. Nasik is currently operating at 80 per cent capacity. The
profit and loss account shows the following:
Rupees in lakhs
Sales 640
Cost of Sales:
Direct materials 200
Direct expenses 80
Variable overheads 40
Fixed overheads 260
580
60
The managing Director has been discussing an offer from Middle East of a quantity
which will require 50 per cent capacity of the factory. The price is 10 per cent less than
the current price in the local market. Order cannot be split. You are asked by him to find
out the most profitable alternative. The factory can be augmented by 10 per cent adding
facilities at an increase of Rs.40 lakhs in fixed cost.

***

65
LESSON – 4
LINEAR PROGRAMMING

Linear programming (LP) is an optimization technique. In business there are


scores of problems requiring optimal solutions. Generally the resources of a business are
of a business are limited. These resources can be put to multiple was. Minimum of
maximum levels for certain uses are fixed, adding further constraints. Under the
situations, the business wants to maximize return, revenue or profit or minimize cost or
consumption of resources. These situations can be effectively handled through the LP
technique. LP technique helps in finding the optimal solution. The minimization /
maximization objective is fulfilled subject to the different constraints.

Linear programming has its origin in the input-output analysis developed by


W.W.Leontiff, a reputed economist. LP assumes that the variables are in linear, i.e.,
straight line, relationship with one another. It assumes that the problem can be stated in
mathematical equations. The term ‘programming’ refers to the use of certain
mathematical techniques to arrive at the optimum solution. ‘Programming’ also means a
sort of routine, or doing things as per predetermined rules and procedures. In the LP
technique a linear objective of the firm and a set of linear equations/inequalities being the
constraints are involved. We have to optimize the objective function satisfying the
constraints.

4.1 CONCEPT OF OPTIMIZATION

Optimization is maximization subject to constraints and minimization to


constraints. Business wants maximization of sales revenue, profit, market share, etc. But
limited resources, multiple uses of resources, price differences of resources, etc. put
difficulties on maximizing sales revenue, profit, etc. So, constrained maximization i.e.,
maximization subject to constraints is sought. This is optimization. In the same way
minimization of cost, time or risk subject to constraints is called constrained
optimization.

4.2 USES OF LP TECHNIQUE IN DECISION MAKING AND CONTROL

There are diverse situations where the LP technique can be adopted. Prominent of
them are enumerated below:

i) In the field of production management the distribution of limited resources (raw


material, labour hours, machine capacity, etc.) to different uses (products,
divisions, markets) so as to maximize contribution is decided using LP
technique.
ii) In production management, determining the optimum mix of inputs so that total
cost of production is minimized, subject to some constraints as to minimum or
maximum limits on the use of one or more inputs.
iii) Instead using the available resource in production of certain goods or services, the
resources themselves can be dispose, if so decided. In such cases the

66
minimum price at which these resources may be sold can be determined
through LP technique.
iv) In marketing management, sales-mix can be decided to maximize profit/revenue
given the capacity constraint for each product and the demand constraint of
each market for the different products.
v) In distribution management to decide the least/cost transportation of goods
from certain number of supply centers, each with certain quantity of the goods,
to certain number of demand centers, each requiring certain quantity, given the
origin-destination-wise transportation cost per unit. (A better method is the
transportation model to be dealt with in the next lesson).
vi) In the case of a banking concern to decide the sector-wise credit-mix so as to
maximize return, subject to overall sectoral limits and region wise sectoral
credit plans.
vii) In financial management to decide the least cost capitalization, subject to ceiling
on the use of on one or more forms of capital.
viii) In farm management to decide the acreage under each/ crop to maximize
return given the return per acre for each crop and constraints as to amount
available for land preparation and the total acreage.
We can go on giving examples like the above. But this is not the
intention. You should note whenever a gain is to be maximized or an outgo is
to be minimized and there are many linear constraints, LP technique can be
used.

PRESENTATION OF PROBLEMS IN LP MODE

In presenting business problems in LP mode certain terminologies are involved.


These are: objective function, constraints, inequality and non-negativity constraints, etc.
These you will know as we present the problems in LP mode.

Illustration 1:
Consider this. A firm makes a contribution to Rs.10 and Rs.12 from the two
products A and B, respectively, it manufactures. One unit of A requires 2 units of raw
material-I and 3 units of material-II, and one unit of B needs 1 unit of material I and 4
units of material II. The firm has 60 units of material I and 120 units of material II. Find
the optimum number of units of A and B.

In this case the objective is to maximize total contribution. Total contribution is a


function of quantity and per unit contribution. If we decide that ‘a’ units of product A
and units of product B, be produced, then the total contribution is given by 10a + 12b.
One unit of A gives Rs.10; therefore ‘a’ units will give 10a. similarly one of B gives
Rs.10; so ‘b’ units will give 12b amount of contribution. So total contribution = 10a +
12b. For a particular value of a and b, 10a + 12b would be the highest. Our objective is
to find those values of a and b. Hence the objective is: maximize 10a + 12b. This is what
we call as objective function, i.e., the function representing our objective.

67
If there were no constraints, as unit contribution is maximum in the case of B, to
maximize total contribution we may simply product ‘B’ only. But there are constraints.
One unit of A requires 2 units of material I and one unit of B requires 1 unit of material I.
We have 60 units of material I and decide to produce ‘a’ units of A. So, 2a units of
material I are required for production of the given level of product A; and the number of
units of material I required to produce ‘’: units of B is lb. Total of 2a + 1b should be
equal to or less than the total of I available, viz. 60 units. That is, 2a + b < 60. Similarly
for material II, 3a + 4b<120. The symbol, < means “equal to or less than”. Further the
number of units of A and B cannot be less than zero. In other words, a and b, can be
equal to or more than zero. This we write as a, b > 0.

So, the constraints are


2a + b < 60 …. (1)
3a + 4b < 120 …. (2)
And a, b > 0 …. (3)

The last one is referred to as the non-negativity condition, since it stresses that the
variable, a and b, cannot take negative values. Items (1) and (2) as well as (3) are not
equations form. The R.H.S. is not equal to L.H.S. Hence these are not equations or
otherwise put these are inequalities. So far the terminologies are explained.

Now we can put the problem as below:

Objective function : Maximize Profit = 10a + 12b


Subject to : 2a + b < 60 │Constraints
3a + 4b < 120 │
a, b > 0 Non-negativity condition.

You may now do some exercises on the mathematical formulation of an LPP.

Illustration 2:

A firm has to produce 400 lbs mixture containing three ingredients A, B and C, a
pound of each costs Rs. 18, 8 and 10 respectively. The total volume of A and C should
not be less than 250 lb, that of B should not be more than 100 lbs and that of C not more
than 125. Find the optimal mix.

Solution:

Here you need to produce a mixture of A, B and C measuring in all 400 lbs. In so
doing you have to minimize cost. And there are other constraints as well. Let us decide
that “a” units of ingredient ‘A’, “b” units of ‘B’ and “c” units of ‘C’ are mixed. Then the

Objective function: minimize Cost = C = 18a + 8b + 10c


Subject to: a + b + c =400

68
a+c > 250
b < 100 Constraints
c < 125
a, b, c > 0 Non-negativity condition.

Illustration 3:

Let us take another example. A form produces 2 products, M and N, 1 unit of M


requires 15 and 20 minutes to grind and assemble respectively, and 1 unit of N needs 15
and 10 minutes respectively. The production run calls for at least 7.5 hours of grinding
time and at least 8 hours of assembly time. If M costs Rs.60 and N Rs.90 to
manufactures find the optimum number of units of M and N.

Solution:

This is a minimization problem. We have to minimize cost. If we take that ‘m’


units of M and ‘n’ units of N are to be produced the objective function would be:
Minimize 60m + 90n

This is to be done subject to:


30m + 30n > 450 minutes │
24m + 48n > 480 minutes │constraints and non-negatively condition
m,n >0
(Note 7 – 5 hours – 450 minutes and 80 hours = 480 minutes).
Since the grinding and assembling hours have to be at least 7.5 hours and 80 hours,
30m + 30n be equal to or more than 15 hours or 900 minutes and 24m + 4bn be equal to
or more than 8 hours or 480 minutes.

Illustration 4:

Take another example. Old hens cost Rs.2 each. Young ones cost Rs.5 each. The
old lays 3 eggs a week and the young 5 a week. Each egg is worth a 30 paise of Re.0.3,
Re.1 is needed to feed a hen in a week. If Rs.80 are available and not more than 20 hens
be maintained at a time, find the optimal mix of old and young hens.

This is a maximization problem. Let us take ‘a’ number of old and ‘b’, number of
young hens. Total egg production per week is therefore 3a + 5b. The value of this at the
rate of 30 paise per egg is: Re.0.3 (3a + 5b). This is gross earnings. We have to feed the
hens which cost Re.1 per hen per week. So the weekly feeding costs Rs.1 (a + b) or a +
b.

So, net earnings per week = .3(3a + 5b) – (a +b). So our objective function is:
Maximize: 0.3(3a + 5b) – (a + b) = 0.9a + 1.5b – a – b = 0.5b – 0.1a
Subject to: 2a + 5b < 80
(the purchase cost be equal to or less than Rs.80)
: a + b < 20

69
(number of hens at any time be restricted to 20 or less)
: a, b > 0 (non-negativity condition).

4.3 USE OF GRAPHICAL METHOD OF SOLVE LPP

So fat we just formulated LPP in a mathematical way. We have to solve them.


Graphical method can be used to solve LPP, whenever the number of unknown variables
is 2. We use certain terms like, feasible solution, bounded and unbounded area here.
These will be explained as we solve the problems.

Illustration 5: Maximization problem through graphic solution

Consider our illustration 1 given earlier. We may solve the problem using the
graphical method. Objective function: Maximize P = 10a + 12b
Subject to: 2a + 5b < 60
3a + 4b < 120
a, b > 0

To solve that problem, first the inequalities are, for the time being, taken as
equalities. That is,
2a + b < 60, becomes 2a + b = 60 and
3a + 4b < 120, becomes 3a + 4b = 120

On the X-axis product ‘A’ and on the Y-axis product ‘B’ are taken. Say, all the
available material I is used to produce product A only. Then 60/2 = 30 units of A can be
produced. If all of material I is used to produce B, 60/1 = 60 units of B can be produced.
We can obtain a straight line connecting these two points: 30 on X-axis and 60 on Y-axis.
This line represents,
2a + b = 60.

Similarly, if all of material II is used to produce A, 120/3 = 40 units can be


produced. If all of material II is used to produce B, 120/4 = 30 units can be produced;
join these points: 40 on X-axis and 30 on Y-axis. This line represents 3a + 4b = 120.
The area OPQR is the bounded area, bounded by both the equations. Graph 4.1 gives the
picture. Any point in the bounded are would meet all the constraints we set in Example
1. So, the bounded area represents all feasible solutions. Feasible Solution thus refers to
any solution meeting all constraints. Outer the bounded area, either or both of the
capacity constraints is/are not met.

The solution to the problem lies at the corner points of the bounded area viz., O, P
and R. Point ‘O’ represents nil production of both A & B. Well that is also feasible
solution as it satisfies all the constraints. But the contribution is zero. Point P means a
production of 30 units of A and no unit of B. The contribution is 10 x 30 + 12 x 0 = 300.
Point R represents 30 units of B

70
and no unit of A. The contribution is 10 x 0 + 12 x 30 = Rs. 360. Point Q can be solved
by solving the two equations:
2a + b = 60 - (1)
3a + 4b = 120 - (2)
Multiply (1) by 4 and we get, 8a + 4b = 240 - (3)
(2) – (3), -5a + 0 = -120
Or, a = 24. Then b = 12. This means 24 units of A and 12 units of B are
represented by point Q.

The contribution is (10 x 24) + (12 x 12) = Rs.384.

Of the 4 corner points, point ‘Q’ produces maximum contribution. Hence 24 units
of A and 12 units of B would have to be produced to maximize contribution.

Alternatively, we choose a minimum contribution level, say Rs.120. To meet this,


12 units of A and 10 units of B are needed. Get a straight line corresponding to 12 units
of A and 10 units of B. Draw parallel lines to the above line. That parallel line farthest
from origin and touching any one of the points indicates the optimal solution.

Illustration 6

Illustration 3 given earlier may be taken up now for solution.


Given: Minimize cost C = 60m + 90n
30m + 30n > 50 min.
24m + 48n > 480 min.
m, n > 0

Now we convert inequalities into equalities. So we get


30m + 30n = 450
24m + 48n = 480

If all grinding time, viz., 450 minutes, is used to produce M, 30 units of M can be
produced. If this time were used to produce N, then a maximum of 360 units can be

71
produced. By joining these two points on X and Y axes, respectively, we get the line
representing 30m + 30n = 450. Similarly if all available assembly time is used to
produce M, 24 units of M, or if diverted to produce N, 48 units of N can be produced. So
we get line, 24m + 48n = 480. You may note from the inequality that 30m + 30n > 450
and 24m + 48n > 480. Graph 4.2 gives the picture.

So, the solution lies in the area outside the bounded area of the two lines on the
graph, 4.2. In other words, the area lying beyond PQR (the shaded area) contains all
feasible solutions. As we want to minimize cost, and also satisfy the inequality
constraints, we would get the solution if we solve for P, Q and R. Corner ‘P’ satisfies all
constraints, i.e., assembling time is 480 mts., and grinding time well over 450 mts. Here
the production cost is Rs. 48 x 90 units of N = Rs. 4320. If we consider corner R,
representing 30 units of M, the cost would be Rs. 30 x 60 = Rs. 1800. This point also
satisfies all constraints. Point Q is solved by solving the two equations.

30m + 30 = 450 – (1)


24m + 48n = 480 – (2)
(1) x 4 we get 120m + 120n = 1800 – (3)
(2) x 5 we get 120m + 240n = 2400 – (4)
(3) - (4), -120n = -600
n = 5; m = 10

Then the cost of 10 M and 5 N = (60 x 10) + (90 x 5) = Rs.1050.


Of the three options, production of 10 units of M and 5 units of N would cost the
least, viz., Rs. 1050. So, this is the optimum production mix.

4.5 USING SIMPLEX METHOD TO SOLVE LPP

LPP can be solved through another method called simplex method. It was
mentioned earlier that graphical method is effectively only when 2 variables are there. If
there are more number of variables, different method is called for. Simplex method is the
other method.

Simplex method is an interactive (repeat) procedure which either solves a LPP in a


finite number of steps or gives an indication that there is an unbounded solution to the

72
given LPP. Here we use some terms like slack variable or slack resource, artificial
variable or resource basic feasible solution, improvement index, optimal column,
intersectional elements, etc. These will be explained as we take up solutions to LPPs.

4.5.1 Maximization through simplex method

The simplex method, the interactive method, is now adopted to solve a


maximization problem.

Illustration: 7

We may now take the illustration 1 we discussed earlier for adopting the simplex
method. You may recollect that the:

Objective function: Maximize 10a + 12b


Subject to: 2a + b < 60 - (1)
3a + 4b < 120 - (2)
a, b > 0

Solution:

To adopt the simplex method, the inequalities (1) and (2) have to be converted into
equalities by introducing the slack variables. You may note that in 2a + b < 60, the LHS
is less than the RHS.

To equate the LHS and RHS, on the lower side we add few quantity, which we call
as slack resource. This may be referred to by S1, S2,… Sn. So the inequality (1) is
converted into an equality as follows: 2a + b + S1 = 60. The S1 simply tells the unutilized
material 1. If all 60 units of I are used S1 will be zero. If some units are left S1 will be
positive quantity. Whether S1 is zero or some positive quantity, the equation 2a + b + S1
= 60 will hold good. Similarly the inequality 3a + 4b < 120 is converted into an equation,
3a + 4b + S2 = 120, S2 being the slack variable, whose value would be zero or more than
zero. In other words, the non-negativity condition would cover both real and slack
variables. While products A and B would give us profits, unutilized material I, i.e., the
slack resource S1 and utilized material II, i.e., the slack resource S2 would not give any.
So, the return from S1 and S2 is zero, each.

In the simplex method the slack variables are also incorporated in the objective
function. So, the objective function is:

Maximize P = 10a + 12b + 0S1 + 0S2


Subject to: 2a + b + S1 = 60. - (1)
3a + 4b + S2 = 120 - (2)
a, b, S1, S2 > 0

73
Equations (1) and (2) are modified in such a way that both contains the other slack
variable also. The coefficient of the ‘other slack variable’ would be put as zero. This
means ‘other slack variable’ is included only to satisfy some structural requirements, but
no value is attached.

So we get, 2a + b + S1 + 0S2 = 60
3a + 4b + 0S1 + S2 = 120

Now, we can proceed with the solution process. Below, initial solution table is
presented. Tableau – 1 gives the same.

Tableau 1: Initial Solution

Explanation of the Table:

Cj row gives the contribution per unit for the real and slack variables. Variables
row gives the variables – real and slack. Cj column gives contribution per unit for those
variable found in the product mix column. Product mix column gives the products (real
and in slack variables) that enter into the solution. Quantity column gives the quantity
produced for each of the product found in the product-mix column, Zj row gives the loss
that would be incurred if one unit of each of the variables in the variable row is produced.
From Cj if we deduct Zj – the net loss / net gain that would be incurred by producing unit
each of the variables obtained.

Now coming to the table, Cj row gives the contribution per unit for a, b, S1, S2
which are Rs.10, Rs.12, Rs.0 and Rs.0. respectively. The variables row gives the
variables a & b (real) and S1 and S2 (slack).

74
The initial solution in a simplex method is a situation where you do not produce
any real variable. So, a and b become zero, giving S1 and S2 values respectively 60 and
120. These S1 and S2 are, therefore, in the product mix and 60 against S1 and 120 against
S2 are placed in the quantity column. S1 and S2 simply represent unutilized resource and,
hence they do not give any return. So, in the column against S1 and S2, their
contributions viz., Rs,0 and Rs.0 respectively are entered. In the S1 row, below each of
the variables, are written the coefficients of the constraint equation, (1). In the S2 row,
below each of the variables, are written the coefficient of the constraint equation (2). The
figures under each variable in the Zj row are the loss they would result if a unit of the
respective variable is produced. Take variable a, i.e., product A. To produce a unit of A,
2 units of S1 and 3 units of S2 (that is 2 units of material – 1 and 3 units of material – II)
are required. But as S1 and S2 are worth nil, by producing one unit of ‘A’ no loss is
incurred, i.e., (2 X 0 + 3 X 0 = 0). To produce a unit of B, that is product B, 1 unit of S1
and 4 units of S2 are required. But the value of these is only 0. Similarly to produce one
unit of S1, one unit of S1 is required and the value is 0 and similarly for S2 it also the value
zero. In the quantity column of Zj row the total profit by producing the products in the
product-mix shown in the quantity column is given. This is simply the sum of respective
quantity times respective contribution. Here it is 60 X 0 + 120 X 0 = 0. Now the Cj – Zj
row giving figures for each of the variable is worked out. It is simply got by subtraction
of the figures in the Zj row from the corresponding figures in the Cj row given at the top
of the table. Cj – Zj is the net profit by making one unit of each of the variables in the
variables row. So, Cj – Zj for variable a is, 10 – 0 = 10. for b is 12 – 0 = 12: for S1 it is 0
– 0 =: and for S2: 0 – 0 = 0. Obviously, where the net gain is more, you would be
attracted there. Here, variable ‘b’ has a maximum of Rs.12 net contribution per unit.
This ‘b’ column with the highest positive Cj – Zj figure is called as the optimum
column.

The figures in the optimum column against each of the variables in the product
mix column are referred to as inter-sectional element. These are bracketed for easy
identification.

Now we have to decide which of these products, A or B, to be produced, for


tableau 1 simply gives the initial and not the final solution. Final solution is reached only
when all figures in the Cj – Zj row either negative or zero. When we decide to produce a
product, that variable will enter into the product mix column and consequently one
variable already there will have to be removed. Which one to enter and which to leave,
the product mix column are decided in the following manner. The variable of the
optimum enters the product mix column. We have to calculate improvement index (I.I)
values for each product of the present product mix by dividing the corresponding quantity
column figure by the corresponding inter-sectional element.

Improvement Index = I.I = Quantity column volume


Inter – sectional element

75
In so doing, if for a row the inter-sectional element is zero or a negative figure, for that
row I.I will not be calculated and the product of that row will not be disturbed. That is,
that will continue to exist.

Now, I.I for S1 row = 60/1 = 60


I.I for S2 row = 120/4 = 30 → (Leaving Variable)

The variable with the lowest (or lower) I.I will be removed from the product and
mix the variable of the optimum column be introduced in the product-mix. In this case
‘b’ will be introduced and S1 will be removed. The S2 row is called leaving or replaced
row. Now the simplex tableau-2 may be worked out.

In the tableau-2 ‘b’ is entered into product-mix column first. The contribution of
Rs.12 per unit is entered in the Cj column.

As per second solution, 30 units of b are produced. This will leave 30 units of S1
unutilized. Hence this appears in the product-mix and quantity columns. ‘b’ row is the
entering row. It is introduced in the place of S2 row of the previous tableau. The figures
for each of the entering row is obtained simply by dividing value of each column, right
from quantity column, of the leaving S2 row of previous tableau by the intersectional
element of that row. So, for ‘b’ row in the tableau 2, the quantity column figure is
obtained by dividing corresponding figure of the leaving S2 row of the previous tableau
i.e., 120/4 = 30. Similarly for ‘a’ variable the figure is ¾; for ‘b’ variable the figure is 4/4
or 1; for S1, 0/4 or 0; for S2 it is ¼. Now ‘b’ row- the entering row is completed. The S1
row continues in the 2nd tableau. But its values are different from those of tableau 1.
Value of each variable of the S1 row in the tableau 2 is to be obtained as follows:

76
Similarly value:
for ‘a’ = 2 – (3/4 X 1) = 5/4
for ‘b’ = 1 – (1 X 1) = 0
for ‘S1’ = 1 – (0 X 1) = 1
for ‘S2’ = 0 – (1/4 X 1) = -1/4

These, 30, 5/4, 0, 1 and -1/4 are written in S1 row of the tableau 2. Now the Zj and
Cj – Zj are worked out in the usual way. In the Cj – Zj row, we find that only one
positive value in the ‘a’ column is found. That is the optimal solution is not still
achieved. So we have to continue our interation. The optimum column is ‘a’ column
now. Therefore in the next solution tableau variable ‘a’ will enter the product mix. To
find which of the existing items in the product-mix will have to be removed, we calculate
I.I values for tableau 2. The intersectional elements are: ¾ and 5/4, which are bracketed
for easy identification.

So, I.I for ‘b’ row = 30/ (3/4) = 40


I.I for ‘S1’ row = 30/ (5/4) = 24

The lower of the two I.I values is 24 and it is learnt that S1 will have to be removed.
So in the tableau 3, given below, ‘a’ variable enters as the entering or replacing row. Of
course ‘b’ row is also there as it is retained.

The values for row ‘a’ are obtained by dividing S2 row figures of the previous
tableau by the intersectional element of that row. The ‘a’ row is the entering row. The
‘b’ row figures in tableau 3 are obtained in the same way as we obtained the values for S1
row of the tableau 2. Zj and Cj – Zj are obtained as usual. Since all values in the Cj – Zj
are either zero or negative, no more further improvement is possible. In other words, the
optimal solution is reached. So, we produce 24 units of A and 12 units of B and we get
Rs.384 as contribution. For all these you may refer to the product mix and quantity

77
columns. You may further note, the same results had been obtained under our graphical
solution to the above problem.

4.5.2 Minimization through simplex method

Illustration 8
We may take up the following example.
Objective function: Minimize: 60P + 120Q
Subject to: 2P + 3Q > 10
P + 4Q > 12
P, Q > 0

Solution:
The inequalities have to be converted into equations. The LHS is greater. So we
subtract slack variables. We get:
2P + 3Q – S1 = 10
P + 4Q – S2 = 12
If P and Q were to be equal to Zero, then, S1 = -10 and S2 = -12. But this will not
be permitted by the non-negativity condition which stress that no variable can take
negative value. So, the above equations have to be modified. We introduce artificial
resources called A1, A2,…An. The artificial variable (resource) is one which is of very
high cost, taken as, great M, and is a substitute for the various resources. If we add these
A’s to LHS, we get over the problem of not fulfilling non-negativity condition. We thus
get:
2P + 3Q – S1 + A1 = 10
P + 4Q – S2 + A2 = 12

Now the objective function can be rewritten as:


Minimize: 60P + 120Q + 0S1 + 0S2 + MA1 + MA2
Subject to: 2P + 3Q – S1 + S2 + A1 + 0A2 = 10
P + 4Q + 0S1 – S2 + 0A1 + A2 = 12
P, Q, S1, S2, A1, A2 > 0

Now the initial solution is obtained in tableau 1. We start with the artificial
variables, if there are any. That is we take up the high cost option first and move to least
cost options thro’ interations. Since in both the constraints, we have artificial resources,
these will enter our product mix. If in any equation there is no artificial resource, the
slack resource of that constraint will enter the solution. Note both A1 & A2 cost M,
assumed very high figure.

78
Simplex Tableau – I: Initial Solution

Cj Mix Qty 60 120 0 0 M M


P Q S1 S2 A1 A2
M A1 10 2 (3) -1 0 1 0
M A2 12 1 (4) 0 -1 0 1
Zj 22M 3M 7M -M -M M M
Cj - Zj - 60 –3M 120-7M M M 0 0

Optimum column

In the case of minimization problem the column with the highest negative Cj – Zj
value is taken as the optimum column. (You may remember in the case of minimization
problem the optimum column is given by the highest positive Cj – Zj value). Variable
‘Q’ has the highest negative value with 120 – 7M. This enters into solution. The
intersectional elements are: for A1 row, it is 3 and for A2 row, it is 4. The respective I.I
values are: 10/3 and 12/4 or 3. The lower of these is 3. So in our next solution A2 is
replaced by Q. Simplex tableau 2 is given below: ‘Q’ row figures in tableau 2 are
obtained by dividing outgoing A2 row figures by the intersectional element: 12/4 = 3, ¼,
4/4 = 1.0, -1/4, 0 and ¼.

Simplex Tableau – 2

Cj Mix Qty 60 120 0 0 M M


P Q S1 S2 A1 A2
120 Q 3 ¼ 1 0 -1/4 0 ¼
M A1 1 5/4 0 -1 ¾ 1 -3/4
Zi M + 360 5/4M+30 120 -M 3/4M -30 M 30-3/4M
Cj-Zj - 30-5/4M 0 M 30- 3/4M 0 1/4M - 30

OPTIMUM COLUMN

A1 row figures in tableau 2 are obtained as follows: (refer formula given


under maximization):
Qty Column: 10 – (3 X 3) = 1
P Column: 2 – (1/4 X 3) = 5/4
Q Column: 3 – (1 X 3) = 0
S1 Column: -1 – (0 X 3) = -1
S2 Column: 0 – (-1/4 X 3) = ¾
A1 Column: 1 – (0 X 3) = 1
A2 Column = 0 – (1/4 X 3) = -3/4

The optimum column is the ‘P’ column with highest Cj – Zj for value of 30 –
5/4M. The I.I value for Q row: 3/ (1/4) = 12 and for A1 row = 1/(5/4) or 4/5. 4/5 is the

79
lower of the two. So, A1 is replaced by P. The next solution is found in tableau 3. ‘P’
row figures are obtained by the dividing outgoing A1 row figures by its intersectional
element. The figures are: 1/(5/4) = 4/5, (5/4) / (5/4) = 1, 0, -1 / (5/4) = -4/5, (3/4)(5/4) =
3/5, 1/(5/4) = 4/5 and (-3/4) / 5/4= -3/5.

Simplex Tableau – 3
Cj Mix Qty 60 120 0 0 M M
P Q S1 S2 A1 A2
60 P 4/5 1 0 -4/5 3/5 4/5 -3/5
120 Q 14/5 0 1 1/5 -2/5 -1/5 2/5
Zj 384 60 120 -24 -12 24 12
Cj-Zj - 0 0 24 12 M - 24 M-12

The figures for Q row in tableau 3 are obtained as follows:


Qty col: 3 – (4/5 X ¼) = 14/5
P col: ¼ - (1 X ¼) = 0
Q col: 1 – (0 X ¼) = 1
S1 col: 0 – (-4/5 X ¼) = 1/5
S2 col: -1/4 – (3/5 X ¼) = -8/20 or -2/5
A1 col: 0 – (4/5 X 1/4) = -1/5
A2 col: ¼ - (-3/5 X ¼) = 8/20 or 2/5

In the Cj – Zj row all values are non-negative. M being very high figure M- 24 and
M-12 are all positive which you should note. Since no value is negative in the Cj-Mj row
the optimum solution is reached.

So the value of P is 4/5 and that of Q is 14/5. The value of the objective function
is: 60 X 4/5 + 120 X 14/5 = 48 + 336 = 384.

4.6 DUALITY

Every linear programming problem has two sides: the primal and the dual. That
is, in maximization problem there is a minimization problem and vice versa. The LLP
as given taken as the primal. Its other side, is the dual. So for a maximization
problem, maximization is primal and its minimization is its dual. You can solve the
primal thro’ its dual. From primal, the dual’s answer can be obtained and from the dual
solution, the solution to its primal can be obtained. So why do you present the dual at
all? the answer is: firstly a problem can be easily solved by its dual than thro’ the primal.
The primal may have more constraints, whereas its dual may not. Then solving the dual
and finding answers to the primal thro’ the dual is good and easy. The opposite situation
will prevail if the dual has many constraints.

80
4.6.1 What is duality?

Consider our illustration 1 where we produce products A & B with the help of
material I and II. We solved that LLP for maximization of contribution. It can be solved
for its dual. If done so, our question would be what minimum price would be acceptable
to us for a unit of material I and material II? It is not always necessary to convert
materials into goods and sell the produced goods. The materials themselves can be
disposed off, it is applicable to all resources. So, what minimum price would be
acceptable to us is the question.

There we say that 2 units of material I and 3 units of material II are required to
produce one unit of ‘a’ which gives Rs.10 contribution. Suppose, one unit of material I is
worth P rupees and that of material II worth Q rupees. If you sell the materials, that
would be committed to a unit of A, then you would expect that 2P + 3Q >= 10. 2P + 3Q
is the value of both the materials spent to produce one unit of A which gives Rs.10
contribution. So to expect that 2P + 3Q is at least equal to Rs.10 is very reasonable.
Similarly, the value of materials spent on one unit of B is, P + 4Q. This should be at least
equal to the contribution from B, viz., Rs.12. That is P + 4Q >= 12. You would be
happy to get the acceptable minimum for both the materials. You have 60 units of I and
120 units of material II. So their value is 60P + 120Q. So, the objective function is:

Minimize 60P + 120Q


Subject to: 2P + 3Q >= 10
P + 4Q >= 12
P, Q >= 0
This we have solved in our illustration 8.
So, illustration 8 is the dual of illustration 7.

4.6.2 Solving dual through primal

I mentioned that from the primal you can solve for the dual. Take our primal-
maximization done earlier. As per its dual, the minimum acceptable price for a unit each
of material I and material II are needed. Look at the maximization tableau 3 in
illustration 7. Under S1 and S2 columns, in the Cj – Zj row we find values -4/5 and
-14/5. Just convert the negative signs into +ve signs and the values are 4/5 and 14/5.
Thus Rs.4/5 is the value of a unit of material I and 14/5 is the value of material II.

4.6.3 Solving Primal through dual

From the dual (minimization) the solution for the primal (maximization) can be
obtained. Look at the minimization tableau 3 of Illustration 8, S1 and S2 columns. In the
Cj – Zj row we get the figures 24 and 12. Now look at the maximization tableau 3 of
Illustration 7, quantity column. Against A there is 24 and against B there is 12. The
same values are found in the dual’s Cj –Zj row under the corresponding slack resources
columns. Hence the primal is solved through dual.

81
You need to know how to write the dual under some situation. Some cases are
presented.

Illustration 9: You are given the primal:


Maximize: 15A + 20B
Subject to: 2A + 4B = 100
5A + 4B =< 160
A, B >= 0
Write the dual.

Solution:
Before we attempt to write dual, the constraints should be converted such that all
of them are of the same form, i.e., “equal to or greater than” or “Equal to or less than”.
The equal to type constraint, 2A + 4B = 100 can be dealt by two inequalities as follows:

2A + 4B >= 100 and - (1)


2A + 4B =< 100 - (2)
Then (1) can be written as
-2A – 4B < -100

Now all the constraints are of the “equal to or less than” form. That is:
-2A – 4B < -100 - (4)
2A + 4B < 100 - (5)
5A + 4B < 160 - (6)

If we take (4), (5) and (6) as different resources with unit value of Rs. P, Q and R
respectively, the dual (minimization) can be written as follows:

Minimize: -100P + 100Q + 160R


Subject to: -2P + 2Q + 5R >= 15
-4P + 4Q + 4R >= 20
P, Q, R >= 0

Note that when the constraints of the primal are “Equal to or less than type”’
those of the dual are “equal to or more than type” and vice versa. Of course non-
negativity condition is always of the “equal to or greater than type”.

Illustration 10:

Jim Jones produces inexpensive furniture to students. Currently it produces book


cases and tables. Each book-case gives a yield of Rs.60 and each table Rs.50. Each
product has to pass through two stages of production, namely, cutting and finishing.
Bookcases take 4 hrs a unit in cutting and 4 hours in finishing. Tables take respectively 3
hrs and 5 hrs. Currently, there are available 96 hrs in cutting and 120 hrs in finishing in
the factor y per week. Solve the LPP graphically and through simplex.

82
Solution:
Let B units of book cases and T units of tables be produced. The yield from book
cases = Rs.60 x B = Rs.60B and yield from table is Rs. 50 X T = Rs. 50T. So, the
objective function is:

Maximize Y = 60B + 50T


Subject to: 4B + 3T < 96 (cutting constraint)
4B + 5T < 120 (finishing constraint)
B, T > 0

Graphic Method
First convert the inequalities into equalities as below:

4B + 3T = 96 (cutting equation)
4B + 5T = 120 (finishing equation)

If all cutting hrs are used to produce book-cases, 24 cases (96/4) can be produced
and if all are used in production, 32 tables (96/3) can be produced. Similarly using all
finishing hrs in bookcases, 30 cases (120/4) and in tables 24 tables (120/5) can be
produced. In the graph 3, book cases are taken in X-axis and tables in Y-axis. The
production possibilities are shown. The feasible solution are a is bounded by OPQR.

The value of Q is got through solving the two equalities:


4B + 3T = 96 … (1)
4B + 5T = 120 … (2)
(1) – (2) -2T = -24
T = 12
So, B = 15

We have to evaluate the corner points, O, P, Q and R


For ‘O’, the yield is: 60 x 0 + 50 x 0 = 0
For ‘P’, the yield is: 60 x 24 + 50 x 0 = 1440
For ‘Q’, the yield is: 60 x 15 + 50 x 12 = 1500
For ‘R’, the yield is: 60 x 0 + 50 x 24 = 1200

So, optimum production is 15 book cases and 12 tables gives a total yield of Rs.1500.

83
Simplex Solution

In simplex solution, we have to introduce slack variables to convert the inequalities


into equalities as below:

Objective function: Maximize Y = 60B + 50T + 0S1 + 0S2


Subject to: 4B + 3T + S1 + 0S2 = 96
4B + 5T + 0S1 + S2 = 1120
B, T, S1, S2 > 0

The initial solution table is worked out now.

The optimum column is ‘B’ (i.e. book cases) and that it enters into production. So
it will be the first now in Tableau – 2. Since, improvement index is low for S1, S2 will be
replaced by B in tableau – 2.

84
The values of ‘B’ row are obtained by dividing the values of S1 row (outgoing row)
by its intersectional element 4. We get: 96/4 = 24, 4/4 = 1, ¾, ¼ and 0.
The values of ‘S2’ row in the tableau -2 are obtained as follows:
Old values in tableau 1 – (Incoming B row values in tableau 2 x S2 row’s intersectional
element in tableau 1)

120 - (24 x 4) 24 (Qty column)


4 - (1 x 4) = 0 (‘B’ column)
5 - (3/4 x 4) = 2 (‘T’ column)
0 - (1/4 x 4) = -1 (S1 column)
1 - (0 x 4) = 1 (S2 column)

Now the optimum column i.e., the column with highest positive number in the Cj –
Zj row, is that of T. So, in our next improved solution, T i.e., Tables, will enter
production mix. When ‘T’ enters, something must be removed from existing product
mix, either B or S2 which has to be removed? This is decided by the lowest value of
improvement index found in the last column of tableau – 2. It is least foe S2. So, S2 will
be outgoing and B will be incoming in tableau 3.

The values of the incoming ‘T’ now in tableau – 3, are obtained by dividing the
values of S2 row in tableau – 2 by its intersectional element. So, the values for ‘T’ row
are: 24/2 = 12, 0, 2/2 = 1, -1/2 and ½.

The values of ‘B’ row in the tableau – 3 are obtained as follows:


Old values in tableau 2 – (Income T row values in tableau 3 x Inter sectional element of
‘B’ row in tableau 2)

24 - (12 x ¾) = 15
1 - (0 x ¾) =1
¾ - (1 x ¾) =0
¼ - (-1/2 x ¾) = 5/8

85
0 - (1/2 x ¾) = -3/8

Since the values in Cj – Zj row in tableau -3 are negative or zero, firther


improvement is not possible. So, 12 units of ‘T’ (tables) and 15 units of B (Book cases)
will be produced giving a total contribution of Rs.1500. All these values are found in
quantity column of tableau – 3. The same result was obtained through graphic solution as
well, as you are aware of.

Duality of the above problem:

In the duality of this problem, we have to find the solution as to what minimum
price for each hour of cutting facility and from each hour of finishing facility will make
the firm earn at least the yield it is getting by manufacturing book cases and tables using
these facilities.

Let an hour of cutting can be hired out for Rs.C and finishing for Rs.F. We have to
find the minimum of ‘C’ and ‘F’.

We now attempt to set the objective function and constraints. We have 96 hrs of
cutting and 120 hrs of finishing. So, 96C + 120F give the objective function.

Minimize: 96C + 120F


Subject to: 4C + 4F > 60 and
3C + 5T > 50
C, T > 0.

That is 4 hrs of cutting and 4 hrs of finishing used in producing 1 book case, must
be hired out such that the value got, i.e., 4C + 4F, is equal to or greater than the yield of
Rs.60, obtained by selling a book-case. Similarly for table usage, 3C + 5T must be equal
to or greater than or equal to yield of Rs.50 obtained from producing one table.

The answer to the minimization duality can be got from the simplex of the primal
maximization solved already. In the final tableau of primal, in the Cj – Zj row, under S1
and S2 columns we find values -12.5 and -2.5. These are the amounts of by which profit
will fall if a unit of S1 and S2, respectively, are created. Now that all units of cutting and
finishing hours are used, creating slack hours will mean, diverting them from production.
So, profit will have to fall. Indirectly, these figures tell the price per hour of cutting
facility and per hour of finishing facility. So, C = Rs.12.5 and F = Rs.2.5. With 96
cutting hours and 120 finishing hours, we get: 96 x 125 + 12.0 x 2.5 = Rs.1500. This is
the same value of objective function obtained for primal as well.

4.7 SENSITIVITY ANALYSIS

Let us take an example. A publishing firm can publish technology book, children
book and business book. Each of these respectively gives a net profit of Rs.2, Rs.4 and
Rs.3.

86
Three resources, authorship, filming and printing are needed. The available
capacities are 60, 40 and 80 hours each, respectively in a week

Technology book requires 3, 2 and 1 hours of each of the resources


Children book requires 4, 1 and 3 hours of each of the resources
Business book requires 2, 2 and 2 hours of each of the resources

If we take ‘T’ units of Tech. Book, ‘C’ units of Children book and ‘B’ units of
Business book be produced, the objective function is:

Maximize : 2T + 4C + 3B
Subject to : 3T + 4C + 2B < 60
2T + 1C + 2B < 40
1T + 3C + 2B < 80
T, C, B > 0

The final solution to the problem is given in tableau – 1. (Initial and improvement
solutions are not given).

Tableau – I: Final solution


Cj Product Qty 2 4 3 0 0 0
Mix T C B SA SF SP

4 C 6 2/3 1/3 1 0 1/3 -1/3 0

3 B 16 2/3 5/6 0 1 -1/6 2/3 0

0 Sp 26 2/3 -5/3 0 0 -2/3 -1/3 1

Zj 76 2/3 23/6 4 3 5/6 2/3 0

Cj – Zj - -11/6 0 0 -5/6 -2/3 0

SA, SF and SP are slack resources as to authorship, filming and printing. The final
solution tells that, 6 2/3 children books and 16 2/3 business books be produced, giving a
total profit of Rs.76 2/3. No slack capacity in authorship and filming is available. But,
26 2/3 hrs of printing is available. As a result printing capacity has no value, (see ‘0’ in
Cj – Zj row), while authorship has a value of Rs.5/6 per hour and filming Rs. 2/3 per
hour. These are known as “Shadow prices”.

Sensitivity analysis is done with respect to right hard side range (RHS range) and
objective function coefficients.

4.7.1 RHS Range

87
Since, excess printing capacity is there, by adding authorship, and filming capacity
we can make additional profit. But we cannot add unlimited authorship and/or filming
capacities as this would affect the constraint equations beyond a point. In other words,
for how large addition of these resources, the ‘shadow point’ will hold good? This is
sensitivity analysis, analyzing the sensitivity of shadow prices remain same, but after that
level these prices remain same, but after that level these prices become sensitivity and
change.

Sensitivity analysis is done as follows. Let us do the sensitivity w. r. t authorship.


For that we reproduce quantity and SA columns of final solution below and divide the
former by the latter row-wise and get the values as follows in tableau – 2.

Tableau – 2: Sensitivity w. r. t. authorship

Quantity column SA column Quotient = Quantity / SA

6 2/3 1/3 (6 2/3) / (1/3) = 20/3 x 3/1 = 20


16 2/3 -1/6 (16 2/3) / (-1/6) = 50/3 x -6/1 = -100
26 2/3 -2/3 (26 2/3) / (-2/3) = 80/3 x -3/2 = -40

The fastest positive quotient, here 20, is the level by which authorship resource can
be reduced without altering shadow prices and least negative quotient, here 40, is the
quantity by which authorship resource can be increased without altering shadow prices.
So, the shadow price for authorship resource, at Rs.5/6, is valid for a range of (-20) to
(+40) hours of authorship hours. Since we have started with 60 hrs of authorship
resource, the right hand side range for it is 40 (i.e. 60 – 20) to 100 (i.e. 60 + 40).

Sensitivity of shadow price of SF: Tableau - 3

Quantity column SA column Quotient = Quantity / SA

6 2/3 -1/3 (6 2/3) / (-1/3) = 20/3 x -3/1 = -20


16 2/3 -2/3 (16 2/3) / (-2/3) = 50/3 x 3/2 x 3/2 = 25
26 2/3 -1/3 (26 2/3) / (-1/3) = 80/3 x -3/1 = -80

We conclude that for filming capacity the range for which shadow price holds at
the present level is 15, (40-25) to 60 (40+20).

With regard to the surplus resource, here printing , the shadow of Re.0, will get
altered only when its availability is reduced by quantity more than the slack quantity, i.e.
below 53 1/3 hrs (80-26 2/3). For quantity > 53 1/3 the shadow price will be same as ‘0’
as now.

4.7.2 Sensitivity w. r. t. objective function coefficients

88
This analysis is done in two types. For variables already in solution and for
variables not in the solution.

4.7.2.1 Variable in solution

Here children book is in the solution. Sensitivity analysis here shows, how large or
how small that the coefficient of children book (i.e. how large or how small the profit per
unit of children book) could become without altering the optimal solution. To get the
answer, we present the Cj – Zj row and Cj row from the final solution and divide the
former by the latter.

Tableau 4: Sensitivity Analysis


Cj – Zj -11/6 0 0 -5/6 -2/3 0
Cj 1/3 1 0 1/3 -1/3 0
(Cj – Zj) / Cj -11/6 x 3/1 0 0 -5/6x3/1 -2/3x-3/1 0
Quotient -5.5 0 0 -2.5 2 0

(*Instead of division by Cj, multiplication by increase of Cj is done)

The least positive quotient sets the limit by which profit can rise without altering
optimal solution. Here it Rs.2. The least negative quotient is the amount by which profit
per children book go down without altering optical solution. Here it is Rs.2.5, so, range
is: RS. 1.5 to 6 [(4-2.5) and (4 + 2)].

Similarly we can find that the range for business book is Rs.2 to Rs.8. That is a
decrease up to Re.1 and increase up to Rs.5 will not change optimal solution.

Variable not in solution

In the final solution, technology book is not in solution. The reason is loss from it
is higher at Rs.23/6 than its profit of Rs.2 per book. To come into solution, profit per
technology book must exceed Rs.23/6. From current profit of Rs.2, it must increase by
Rs.11/6 (i.e Rs.23/6 – Rs.2), so that it can enter the solution.

4.8 OTHER TYPES OF LINEAR PROGRAMMING

Production smoothening application, portfolio selection ingredient mixing, diet


combination, etc. are certain special applications of LP.
Transportation and assignment problems are special types of linear programming.
These are dealt in the next lesson.

4.9 LIMITATIONS OF LP

i) Assumption of linearity as to the relationship among variables is a mathematical


convenience, but devoid of practical relevance.

89
ii) The graphic solution seems easy, but cannot be adopted when there are more than
two variables.
iii) The simplex method involves some complications. Those with lack of a
mathematical bent of mind would find it difficult to understand.
iv) LP would become more and more involved when more constraints are there.
Better special purpose algorithm have been developed, thus relegating simplex
method.
v) Degeneracy might result. That is some +ve value in the max. LPP or –ve value
in the min-LPP might continue to appear in the Cj – Zj row, yet the Qty. Col.
Total remains same, indicating no improvement after successive iterations.
vi) Unboundness is another problem. There is no definite solution to the LPP.
vii) Redundancy is a problem, wherein constraints that not at all arise are stipulated.

The last few are however, strictly not limitations. The LPP is not well formulated.
So, it is futile to blame the technique as such.

4.10 CERTAIN CONCEPTS

Infeasibility, unboundedness, redundancy, multi-optimal, degeneracy, etc. are


certain concepts associated with LPPs.

Infeasibility means there is no solution to the LPP meeting all constraints. In


graphic solution, infeasibility is expressed when there is no area of feasible solution. In
simplex solution when one or more artificial resources are found in final solution there is
infeasibility.

Un-boundedness means that in an LPP infinite solutions exists. Practical world is


beset with constraints. So when an LPP produces unlimited solutions, there is something
wrong in its formulation. In simplex method, unboundedness is expressed when the
improvement indices are all negative or zero. This means an infinite result.

Redundancy means, that in the LPP one or more constraints are really no
constraints. When maximum production possible is say only 15 table and 12 chairs, a
marketing constraint of no more than 20 tables and 15 chairs could be sold is a redundant
constraint not affecting the production feasibility at all.

Degeneracy in LPP arises when improvement index values for two or more
variables are equal. Then, one of them is randomly replaced and this leads to degeneracy.
Whenever a variable in the solution has the value 0 in the quantity column, that variable
and that solution are said to be degenerate. When there is a degeneracy, the value of the
objective function does not improve with the next solution.

Multiple optima in an LPP means that the LPP has many solutions giving the same
objective function value.

SELF-ASSESSMENT QUESTIONS

90
1. Explain the LP technique as an optimization technique.
2. What are the managerial applications of LP?
3. Enumerate the limitations of LP.
4. A company produces 3 products A, B and C giving a profit of Rs.1000, Rs.4000
and Rs.5000 per unit respectively. O1, O2 and O3 are the materials required.
The requirement is as follows per unit of output of A, B and C:
- A B C
O1 3 - 3
O2 1 2 3
O3 3 2 -
The available quantity of 01, O2 and O3 is 22, 14 and 14 units respectively. To
maximize profit, find the optimum product mix.

5. Write the dual for the above problem, solve the same.
6. Maximize: 6X + 4Y
Subject to: 2X + 3Y < 30
3X + 2Y < 24
X+Y>3
X, Y > 0.
7. Writes notes on:
a) Initial solution; b) Objective function; c) Non-negativity condition; d)
Constraints; e) Feasible region; f) Optimum column; g) Intersectional
elements; h) Artificial resource
8. What are degeneracy, redundancy, unboundedness and infeasibility?
9. Explain the significance and process of sensitivity analysis.
10. A Toffee company mixes 3 types of toffees to form one kilogram of Toffee gift
pack sold at Rs.300. The three toffees cost Rs.200, Rs.250 and Rs.280 per kg
each. The mixture must contain at least 0.3 kg of first type of Toffee and the
weight of the first two must at least be equal to that of the third. Find the optimal
mix.

***
LESSON 5

TRANSPORTATION AND ASSIGNMENT MODELS


(Special Purpose Linear programming)

Transportation and assignment models are special purpose algorithms of the linear
programming. The simplex method of LPP seen in the previous lesson proves to be
inefficient is certain situations like determining optimum assignment of jobs to persons,
supply of materials from several supply points to several destinations and the like. More
effective solution models have been evolved. These are called assignment and
transportation models which are dealt with in this chapter.

5.1 MEANING:

91
The transportation model is concerned with selecting the routes between supply
and demand points in order to minimize costs of transportation subject to constraints of
supply at any supply point and demand at any demand point. Assume a company has 4
manufacturing plants with different capacity levels, and 5 regional distribution centres.
4 x 5 = 20 routes are possible. Given the transportation costs per load of each of 20
routes between the manufacturing (supply) plants and the regional distribution (demand)
centres, and supply and demand constraints, how many loads can be transported through
different routes so as to minimize transportation costs? The answer to this question is
obtained easily through the transportation algorithm. Similarly, how are we to assign
different jobs to different persons/machines, given cost of job completion for each pair of
job machine/person? The objective is minimizing total cost. This is best solved thro’
assignment algorithm.

5.2 USES OF TRANSPORTATION AND ASSIGNMENT MODELS IN


DECSION MAKING:

The broad purposes of these models are just mentioned above. Now we have just
enumerated the different situations where we can make use of these models.
i) To decide the transportation of new materials from various centres to different
manufacturing plants. In the case of multi-plant company this is highly
useful.
ii) To decide the transportation of finished goods from different manufacturing
plants to the different distribution centres. For a multi-plant-multi-market
company this is useful.
iii) To decide the transportation of finished goods from different manufacturing
plants to the different distribution centres. For a multi-plant-multi-market
company this is useful. These two are the uses of transportation model. The
objective is minimizing transportation cost.

Assignment model is used in the following:


i) To decide the assignment of jobs to persons/machines, the assignment model is
used.
ii) To decide the route a traveling executive has to adopt (dealing with the order inn
which he/she has to visit different places).
iii) To decide the order in which different activities performed on one and the same
facility be taken up.

In the case of transportation model, the supply quantity may be less or more than
the demand. Similarly the assignment model, the number of jobs may be equal to, less or
more than the number of machines/persons available. In all these cases the simplex
method of LPP can be adopted, but transportation and assignment models are more
effective, less time consuming and easier than the LPP. Further, transportation and
assignment models are useful only in these specific instances dealt within this lesson.

92
Solution to transportation problems consist of initial solution and improved
solutions. To find initial solution North West Corner Rule Methods, Vogel’s
approximation, or row cost minima, etc, are used. For improved solution Stepping stone
and ‘MODI’ methods are also used.

5.3 MECHANISM OF TRANSPOTATION MODEL:

The mechanism of transportation model is explained with an example.

Illustration 1:

Let us take an example. A construction company has 3 projects, A, B and C.


Truck load of sands required per week for these 3projects are: 80, 100 and 120. Three
supply points of sand are there namely, X, Y and Z with a capacity per week of 120,80
and 100 loads respectively. The delivery cost per load from each supply site to each
demand point are as follows:

Cost per load


To A To B To C
From X 12 8 18
From Y 17 10 12
From Z 19 5 8

Decide how many truck loads of sand from the supply points to the different
projects in order to reduce transportation cost.

Solution
The above problem can be presented in a table form.

Initial solution: North West corner Rule:


Here supply = demand. So this is a balanced problem.

To solve the problem an initial solution has to be worked out. Later this solution is
improved. Initial solution is the starting point. There are several methods to find initial
solution. North west corner method, Vogel’s Approximation method, Row Minima
Method, Column Minima method and inspection method are some of the methods. Let
us take up the North West Corner method to develop the initial solution. Consider table 1
given below.

Project A Project B Project C Total Supply


Supply X XA XB ( XC 120

93
(12 8 (18
(80) (40)
Supply Y YA (17 YB (10 YC (12 80
(60) (20)
Supply Z ZA ZB ( ZC ( 100
(19 5 (100)
Total Demand 80 100 120 300

The table simply reproduces the problem with additional information. The cost
figures are placed in sub-cells within each cell. The XA, XB,….XC simply indicate the
origin – destination pair.

The north west corner rule is logical and systematic method to find the initial
solution. You start with north west corner cell, viz, XA cell. It means sending supplies
from X to project A. X has a total capacity of 120 loads. A needs only 80 loads. So,
‘A’s demand is fully met from X. This allotment is written in a bracket within XA cell.
Since A’s need is fully met we can move to project B. Now balance available with ‘X’
needs to be exhausted first. So we move horizontally from XA cell to XB cell. The
balance 40 loads of X is allotted to B. This is indicated in the bracketed figure in XB
cell. B requires a total of 100 loads. So, it still needs 60 more loads of sands. We have
to use the supply with the next source viz. Y. So we move vertically down to YB from
XB. From Y, 60 loads are to be sent B. This is indicated in the bracketed figure in the
bracket in YB. Now that B’s requirement is fully met, you move to C and still some
capacity is available with Y. Our move is horizontal to YC. 20 units of available balance
with ‘Y’ allotted to project C. Still project ‘C’ needs 100 more loads. We move down to
ZC and the available stock with Z is just sufficient to meet the balance needs of C. This
is a case of supply being equal to demand type problem. Initial solution as obtained
under north west corner method is as follows: From X to A 80 loads, X to B 40 loads, Y
to B 60 loads, from Y to C 20 loads and from Z to C 100 loads are transported. This is
only an initial solution. A better solution you have to find from the above. Arrow marks
are made to present flow of allotments made.

To produce further we have to consider the number of stone cells. Stone cells are
those with allotment. XA, XB, YB, YC and ZC are stone cells. The number of stone
cells must be equal to m + n – 1, where ‘m’ indicates the total number of supply sources
(here 3) and ‘n’ indicates the total number of demand points (here the project numbering
3). 3 + 3 – 1 = 5 = No. of stone cells. This equality is referred to as rim-requirement.
The satisfaction of rim-requirement, i.e., m + n – 1 = No. of stone cells is must for further
attempt on improving the solution at any stage. Now the rim-requirement is met, we can
proceed further.

To proceed hence forth, we have to find what would happen to cost if one load is
allotted to each of the unoccupied cells, or non-stone cells. There are two methods here,
viz. stepping stone method or modified or MODI method. Now we are adopting the
stepping stone method.

94
Stepping Stone method for optimization:

If one load is allotted to XC, one load must be subtracted from XB, one load must
be added to YB and one load must be subtracted from YC. This is necessary so that the
supply constraints of X and Y and the demand levels of B and C are not affected. What
would be the effect on cost of this more? X to C costs Rs.18 and Y to B costs Rs.10.
Withdrawal of a load from X to B saves Rs.8 and form Y to C saves Rs.12. the overall
effect of this: 18 – 8 + 10 – 12 = Rs.8. Sending a load of sand to XC results in extra cost
of Rs.8. this has to be avoided. Similarly for other unoccupied cells evaluation is to be
done. To send a load to YA: YA – YB + XB – XA. The effect on cost would be : 17 –
10 + 8 – 12 = 5. This is also no good.

To send a load A to Z, would require one load less for XA, one load more for XB,
one load less for YB, one load less for ZC. That is: ZA – ZC + YC + YB + XB – XA.
You might note the mechanism. You start from the empty cell (giving a+). You move
either vertically or horizontally, never diagonally to a stone cell. You take turn only at
the stone cell. And so on. Ultimately you get back to the empty cell you started from.
These routes are explained in table 2.

Table 2. Routes to unoccupied cells in table 1.


A B C Total supply
X (80) (40) + 120

Y (60) (20) 80
Z (+) (100) 100
(-)
Total demand 80 100 120 300

The process is one of making ‘loop’, a closed loop. Successive turn points give ‘-‘
and ‘+’ signs alternatively the dash-dash route in table 2 gives this, make sure you never
halt or turn in any intervening cells. The effect on cost of the above more is: 19 – 12 + 8
– 10 + 12 -8 = 9. Again this move is not beneficial as cost goes up by 9 for every load
thus transported. The only remaining empty cell is ZB. For this a simple loop is possible
and the same is: ZB – ZC + YC – YB. The cost effect is: 5 – 8 + 12 – 10 = -1. So,
sending one load to ZB reduces the cost Re.1. This route has to be utilized to the
maximum extent, so that maximum reduction in cost is possible. So we have to do
reallocations among the involved cells here. There are four cells involved. Of these 4
two are ‘+’, i.e., receiving cells and two are ‘-‘, i.e., sending cells. How much to send?
Of the sending cells, find the one with least allocation. That much will be moved. Here
ZC and YB are sending cells. ZC has 100 and YB has 60. So, 60 loads will be moved
from YB to ZB, 60 from ZC to YC. The relevant reallocation would be: ZB: 60 loads:
YB: nil; YC 80 loads and ZC 40 loads. (refer table 3) Note allocations to XA and XB are
not changed. Because these cells, are not involved in the ‘loop’ for ZB. Now also the
rim-requirement is fulfilled as, m + n – 1 = stone cells. So, we can think of further
improvements, if any. Now the empty cells are: XC, YA, YB and ZA. The routes to
these empty cells and additional cost of allocations empty cells would be as follows:

95
XC: XC – XB +ZB – ZC = 18 – 8 + 5 – 8 =3
YA: YA – XA – XB + ZB – ZC +YC= 17 – 12 + 8 -5 + 8 – 12 = 4 (loop
again)
YB: YB – YC + ZC – ZB = 10 – 12 + 8 – 5 =1
ZA: ZA – XA + XB – ZB = 19 – 12 + 8 – 5 = 10

Table 3: Second Solution


A B C Total supply
X 12 8 18 120
(80) (40)
Y 17 10 12 80
(80)
Z 19 5 8 100
(60) (40)
Total demand 80 100 120 300

Allocation to any of the empty cells proves to be inefficient. That is, allocation
them involve additional cost. So, a final solution is already obtained. As per table 3, to
XA 80 loads, XB 40 loads, YC 80 loads, to ZB 60 loads and to ZC 40 loads allocations
are made. The total transportation cost would be:

XA : 8 x 12 = 960
XB : 40 x 8 = 320
YC : 80 x 12 = 960
ZB : 60 x 5 = 300
ZC : 40 x 8 = 320

Total Rs. 2860

Any other allocation would be costlier than the above.

Vogel’s Approximation Method (VAM) for Initial solution:

Vogel’s Approximation Method can also be used, instead of north west corner rule
method to find the initial solution. Under this method the following steps are involved

i) Calculate the difference between lowest of two transportation costs for each
column and each row. These differences are known as column and row penalties
and are written by respective row and column by providing additional column and
additional rows in the transportation table.

ii) Suppose in any row or column the smallest cost figures repeat. These repeat
values are taken for penalty computation and the penalty is case in such cases.

96
iii) Select the row or column with the largest penalty and circle that value. In case of
a tie, select the column or row that permits the greatest movement of materials.

iv) Assign the largest possible allocation within the restrictions of the row and
column requirements to the lowest cost cell for the row or column selected in step
(iii) above.

v) Cross out the column or row whose demand/supply is completely met


with/exhausted, as the case may be, as per our assignment made in the previous
step.

vi) With a column or row thus deleted, (as per step in above) with the reduced
number of columns/rows repeat the process (i) to (ii) until all assignments are
made.

Let us consider our problem in question.

Illustration 2:

Illustration 1 is solved through VAM and MODI method. Table 1 gives the
problem and application of VAM.
Table 1: VAM: PENALTY
A B C Supply Row penalty
X 1 (8) (18) 120 12 – 8 = 4
2
(80)
Y (17) (10) (12) 80 12 – 10 = 2
Z (19) (5) (80) 100 8–5=3
Demand 80 100 120 300
Column 17 – 12 = 5 8–5=3 12 – 8 = 4
penalty

Look at the table 1, the largest penalty is 5 in column A. The least cost cell in that
column is XA, X has 120, but A needs only 80. So, 80 units given to A from X. This
reduces supply with X to 40 and A gets deleted. Now Column A is eliminated as its need
is fully met. So, the reduced matrix is in table 2 below.

B C Supply Penalty
X (40) (8) (18) 40 18 – 8= 10
Y (10) (12) 80 12 – 10 = 2
Z (5) (8) 100 8–5=3
Demand 100 120 220
Penalty 8–5=3 12 – 8 = 4

97
Now the highest penalty is 10 in row X. The least cost cell in row X is XB. 40
units with X are given to B and B > read reduces to 60 from 100 new penalty is now
constructed in table 3 below.

B C Supply Penalty
Y (60) (20) 80 12 – 10 = 2
(10) (12)
Z (5) (100) (3) 100 5–3=2
Demand 60 120 180
Penalty 10 – 5 = 5 12 – 3 = 9

The highest penalty is 9 in column C. The least cost cell is ZC. So, 100 units from
Z are sent to C. Now Z row gets eliminated. This leaves with us only Y row. So,
whatever is available with Y has to be given to ‘B’ and ‘C’ B gets 60 from Y and C gets
20. And this ends all initial distribution

The rim-requirement = m + n – 1 = 5 = Number of stone cells, is fulfilled; so we


can proceed further.

Note this initial solution is same as the one obtained under the North west corner
rule. But of need not be so. Improved solutions have to be found. Table 4 gives the
picture.

MODI method for optimization

At this stage we may adopt the modified or MODI method for improved solution.
Under this method the columns and rows are represented by K’s and R’s respectively.

So, K1, K2 and K3 and R1, R2 and R3 are the notations for the three columns and
three rows respectively. The cost of a cell can be represented by K’s and R’s
respectively.

Cost at Stone (i.e. occupied) cell i, j = Cij = Ri + Kj. In our case under the Vogel’s
approximation method the stone cells are: 1.1, 1.2, 2.2, 2.3 & 3.3. So

C1, 1 = R1 + K1 = 12 - (1)
C1, 2 = R1 + K2 = 8 - (2)
C2, 2 = R2 + K2 = 10 - (3)
C2, 3 = R2 + K3 = 12 - (4)
C3, 3 = R3 + K3 = 8 - (5)

We have 5 equations, but 6 unknowns. So, one of the unknown be given value
zero.

Let any one of Ri’s or Kj’s be equal to zero. Say K1 = 0. Then, as per equation

98
(1) 12 = R1 + 0 or R1 = 12
(2) 8 = 12 + K2 or K2 = -4
(3) 10 = R2 + (-4) or R2 = 14
(4) 12 = 14 + K3 or K3 = -2
(5) 8 = R3 + (-2) or R3 = 10

Now the cost effect of sending one load to each of the unoccupied cells viz., 1,3;
2,1; 3,1; and 3,2; have to be found. This is given by Cij – Ri – Rj

So the cost effect for XC = C13 – R1 – K3


= 18 – 12 – (-2) = 8

So the cost effect for YB = C32 – R2 – K2


= 17 – 14 – (0) = 3

So the cost effect for ZA = C31 – R3 – K1


= 19 – 10 – 0 = 9

So the cost effect for ZB = C32 – R3 – K2


= 5 – 10 – (-4) = -1

Now you look for the cell with the largest –ve cost effect. It is ZB. We have to send
loads from Z to B. the closed loop would be: ZB – YB + YC – ZC.
Now, from YB to YC 60 units are transferred making the total to 80. From ZC to
ZB, 60 units are transferred. The new position is given in table 4.
A B C Total
supply
X 12 8 18 120
(80) (40)
Y 17 10 12 80
(60) (20)
Z 19 5 8 100
(100)
Total 80 100 120 300
demand
Again the cost effect of sending a load to the empty cells, need to be worked out.
For this purpose we must know the values of Kj’s and Ri’s.
Cost at a stone cell: Cij = Ri + Kj
So for, XA: C11 = 12 = R1 + K1 - (1)
XB: C12 = 8 = R1 + K2 - (2)
YC: C23 = 12 = R2 + K3 - (3)
ZB : C32 = 5 = R3 + K2 - (4)
ZC: C33 = 8 = R3 + K3 - (5)

99
Setting any of the Rj’s or Kj’s as equal to zero, we can solve the remaining Ki’s
and Kj’s. Let R1 = 0.

Then 12 = 0 + K1; or K1= 12 (based on equation. 1)


8 = 0 + K2; or K2 = 8 (based on equation. 2)
5 = R3 + 8; or R3 = -3 (based on equation. 4)
8 = -3 + K3; or K3 = 11 (based on equation. 5)
12 = R2 + 11; or R2 = 1 (based on equation. 3)

The cost effect for non-stone cells be worked out using the formula: Cij – Ri – kj

Cost effect for XC: 18 – R1 – K3 = 18 – 0 – 11 = 7


Cost effect for YA: 17 – R2 – K1 = 17 – 1 – 12 = 4
Cost effect for YB: 10 – R2 – K2 = 10 – 1 – 8 = 1
Cost effect for ZA: 19 – R3 – K1 = 19 – (-3) – 12 = 10

Now all figures are +ve indicating no improvement is possible hence form.
Table 5
A B C Total
supply
X 12 8 18 120
(80) (40)
Y 17 10 12 80
(80)
Z 19 5 8 100
(60) (40)

Total 80 100 120 300


demand

So, we have reached the final solution as in table 5. This solution is the same as
the one obtained earlier.

5.4 TRANSPORTATION MODEL WHEN DEMAND IS LESS THAN SUPPLY

This is called unbalanced problem. To solve the problem, first balancing is needed.

Here you add a ‘Dummy’ demand absorbing the excess supply. The cost figures
for this ‘dummy’ demand column would be zero for all rows. Any supply centre
supplying to this ‘Dummy’ demand does not, in fact supply to this destination as the
latter is non-existent. A problem on this pattern is worked out now.

Illustration 3
Given in the table are the supply and the demand factors and the transportation cost
matrix. Find the optimal distribution.
Deport A B C D Total

100
Factory supply
P 4 6 8 6 700
Q 3 5 2 5 550
R 3 9 6 5 550
Total 1800
Need 400 450 350 500 1700

Solution

Here the supply exceeds demand. So a Dummy Depot is to be added with a need
or demand of 100 units. The initial solution under North West Corner Rule is given in
table 1.

A B C D Dummy Total
supply
P 4 6 8 6 0 700
(400) (300)
Q 3 5 2 5 0 550
(150) (350) (50)
R 3 9 6 5 0 550
(450) 100
Total 400 450 350 500 100 1800
demand

Now the rim requirement is m + n – 1 = 3 + 5 – 1 = 7 = number of stone cells. We


can proceed further. The effect on cost or transportation of one unit to each of the non-
stone cells may be worked out now.

For PC: PC – PB + QB – QC = 8 – 6 + 5 – 2 = 5
For PD: PD = PB – QB – QD = 6 - 6 + 5 – 5 = 0
For P Dummy: P – Dummy – PB + QB – QD + RD – R – Dummy
: 0 - + 5 – 5 + 5 – 0 = -1

For QA: QA – PA + PB – QB = 3 – 4 + 6 – 5 = 0
For RA: RA – PA + PB – QB + QD – RD
:3–4+6–5+5–5=0

For Q – Dummy: Q–Dummy – R–Dummy + RD – QD = 0 – 0 + 5 – 5


=0
For RB: RB – QB + QD – RD = 9 – 5 + 5 – 5 = 4
For RC: RC – QC + QD – RD = 6 – 2 + 5 – 5 = 4

From the above cost effect computations, cost minimization is possible only is
distribution from P to Dummy is made.

101
Note that the movement in any improvement is limited to the minimum quantity
of the –ve stone cells in the loop. In this case the loop is P-Dummy- PB + QB – QD +
PD – R-Dummy. QD in old assignment has the minimum of 50 units. So the movement
is limited to successive add/less of 50 units. Any higher movement would result in – ve
figure at the minimum –ve stone cells which s not realistic. Table 2 gives the 2nd
solution.

Table 2:2nd Solution


A B C D Dummy Total
supply
P 4 6 8 6 0
(400) (250) (50) 700
Q 3 5 2 5 0
200 (350) 550
R 3 9 6 5 0
(400) (50) 550
Total
demand 400 450 350 500 100 1800

Now improvement may be worked out for the empty cells in the table 2.

For PC: PC – PB + QB – QC = 8 – 6 + 5 – 2 = 5
For PD: PD – P-Dummy + R-Dummy – RD = 6 – 0 + 0 – 5 = 1
For QD: QD – QB + PB – P-Dummy + R-Dummy – RD = 5 – 5 + 6 – 0 + 0 – 5 = 1
For Q-Dummy: Q-Dummy – QB + PB – P – Dummy = 0 – 5 + 6 – 0 = 1
For RA: RA – R–Dummy + P-Dummy – PA = 3 – 0 + 0- 4 = -1
For RB: RB – R-Dummy + P-Dummy – PB = 9 – 0 + 0 – 6 = 3
For RC: RC – R-Dummy + P-Dummy – PB + QB – QC = 6 – 0 + 0 – 6 + 5 – 2 = 3
For QA: QA – PA + PB – QB = 3 – 4 + 6 – 5 = 0
For RA cell there is –ve cost impact figure.

So, distribution from R to A is advantageous. Table 3 gives the new distribution.

Table 3:3rd Solution


A B C D Dummy Total
Supply
P 4 6 8 6 6
(350) (250) (100) 700
Q 3 5 2 5 0
200 (350) 550
R 3 9 6 5 0

102
(50) (500) 550
Total
demand 400 450 350 500 100 1800

Now, all the empty cells have (computations not given) positive cost effects. So,
table 3 gives the optimum solution. As per this, the distribution is:

P to A: 350 units; cost: 350 x 4 = 1400


P to B: 250 units; cost: 250 x 6 = 1500
(P to Dummy is in effect no distribution) = 0
Q to B: 200 units; cost: 350 x 2 = 700
R to A: 50 units; cost: 50 x 3 = 150
R to D: 500 units; cost: 500 x 5 = 2500
Total Cost Rs. 7250

You may not the transportation as per table 1 involves a cost of Rs.7350 and table
2 involves a cost of Rs.7300. So, successive iterations lead to reduced cost. The
optimum is reached when no further reduction in cost is possible.

5.5 TRANSPORTATION MODEL WITH SUPPLY LESS THAN DEMAND

When supply falls short of demand a dummy is added to the supply source and the
cost transportation from this dummy source to any of the demand centres is taken as 0.
Dummy’s supply is taken as the excess of demand over available supply. The problem is
solved in the usual way. Any destination getting allotment from ‘dummy’ in effect does
not get real allotment to that extent.

5.6 MAXIMIZATION PROBLEM IN TRANSPORTATION

Transportation model is designed to solve cost minimization problems. However it


can be used to solve maximization problems too. Consider the problem.

Illustration 4

A firm has 3 factories, viz., A, B & C. It has 3 sales territories, viz., X, Y & Z.
Excluding transportation cost per unit returns, i.e., the ex-factory returns are, Rs.10, 9 and
11 at these three factories. The transportation costs and capacities factors are as under.

From To X To Y To Z Total supply


A 1 2 3 1000
B 5 3 4 800
C 3 5 3 1200
800 1200 1000 3000

You are required to distribute in such a way, that maximum return is obtained.

103
Solution
Step I
The above cost figures are not to be considered in isolation. First we have to
ascertain the per unit return net of transportation cost. This is as follows: Net return =
Gross ex-factory return – Relevant transport cost. This is given below in the table 1.

Table 1: Net Return Matrix


X Y Z
A 10 – 1 =9 10 – 2 = 8 10 – 3 = 7
B 9–5=4 9–3=6 9–4=5
C 11 – 3 =8 11 – 5 = 6 11 – 3 = 8

Step II:
This return per unit matrix has to be converted into a cost matrix. This is done as
follows. Find the highest net return figure. It is Rs. 9. Subtract all net return figures
from 9 and construct the matrix. This is done in the table 2 that follows, which also gives
the quantity factors and initial solution.

Table – 2 – Initial Solution: NWCR Method


X Y Z Total
Supply
A 0 1 2
(800) (200) 1000
B 5 3 4
(800) 800
C 1 3 1
(200) (1000) 1200
Total
Demand 800 1200 1000 3000

Table – II – Final Solution

X Y Z Total
Supply
A 0 1 8
(600) (400) 1000
B 5 3 4
(800) 800

104
C 1 3 1
(200) (1000) 1200
Total
Demand 800 1200 1000 3000

Table 2 solution is the optimum one. Detailed workings are not given. So, from A
600 units to X and 400 units to Y, from B 800 units to Y, from C 200 units to X and 1000
units to Z are to be sent. The profit would be = units times net return.

= 600 x 9 + 400 x 8 + 800 x 6 + 200 x 8 + 1000 x 8 = Rs.23000

5.7 ASSIGNMENT MODEL

Assignment model is a special case of transportation problem. Assignment of jobs


to machines/persons, sequence in which jobs be undertaken, sequence in which places
have to be visited etc., are better determined using assignment model. In all these cases
cost minimization is the root objective. Then one-to-one basis underlies the assignment.
That is, one machine to one job, is the basis. Similarly, in sequence determination, only
once a job or place be taken up/visited and no retreat is permitted. Let us examine the
use of the assignment model now. The method adopted here is called Flood’s technique
or the Hungarian method of assignment.

JOB-MACHINE ASSIGNMENT

Illustration 5:

The cost matrix for each job-machine combination is as follows:


Job Mac. X Mac. Y Mac. Z
A 30 30 25
B 45 18 25
C 21 17 15

Find the optimum assignment of jobs to machines.


Step 1: Column-wise subtract the least cost figure from the respective row figures.
This is done as found in table 1.

Table 1
Job Mac. X Mac. Y Mac. Z
A 30 – 21 = 9 30 – 17 = 13 25 – 15 = 10
B 45 – 21 = 24 18 – 17 = 1 25 – 15 = 10
C 21 – 21 = 0 17 – 17 = 0 15 – 15 = 0

Step 2: For resulting figures, go row-wise subtracting each row minimum from
respective row figures. We get table 2 as follows:

105
TABLE 2
Job Mac. X Mac. Y Mac. Z
A 9 – 9 = 0 13 – 4 = 9 10 – 9 = 1
B 24 – 1 = 23 1 – 1 =0 10 – 1 = 9
C 0 0 0

Step 3: Draw straight lines, to connect the zeros in the above table. Minimum number of
lines must be used. If that minimum number of lines equals the jobs/machines, an
assignment is possible. Here it is possible.

Step 4: Match jobs with machines wherever the cell figures are zero. Give priority to
row/column with only one zero. Z column has only one zero. So, job C be assigned to Z;
and the zero is CZ is boxed. Any other zero in C row to Z column will be crossed out.
This reduces the number of zeros in other rows and columns too. Row A has only one
zero; in X column. So, job A be assigned to machine X. This leaves job B be assigned to
machine Y.

The assignment is therefore A to X, B to Y and C to Z. The cost = 30 + 18 +15 =


63.

Illustration 6:
The job-machine cost matrix is given below:
Job Mac. X Mac. Y Mac. Z
A 25 31 35
B 15 20 24
C 22 19 17
Determine the optimum assignment

Solution:

Step 1: Row-wise, subtract minimum value from respective row figures.


(instead column wise, you can do this way also). 25 is subtracted from row 1 figures, 15
subtracted from row 2 figures and 17 from row 3 figures.

We get figures as found in table 1.

Table 1
Job Mac. X Mac. Y Mac. Z
A 0 6 10
B 0 5 9
C 5 2 0

106
Step 2: for the resulting figures we go column-wise deducting each
column minimum value from respective column figures and we get figures as found in
table 2.
Table 2
Job Mac. X Mac. Y Mac. Z
A 0 4 10
B 0 3 9
C 5 0 0

Step 3: Draw St. lines to connect the zeros. We need only a minimum of
two st. lines for the purpose. Since the number of st. lines is less than the number of
jobs/machines, optimum assignment is not decided at this stage. In such case we proceed
further

Step 4: Find the lowest value number not covered by any st. line.
Subtract this number from each of the numbers not covered by any st. line. Add that
number to the figure found at the intersection of any two st. lines. Other numbers are left
untouched. Accordingly, the lowest number is 3. This is subtracted from 4, 10, 3 and 9
and added to 5 at the intersection. We get:

Job Mac. X Mac. Y Mac. Z


A 0 1 7
B 0 0 6
C 8 0 0

Now draw lines to connect the zeros. We need a minimum of 3 lines which is
equal to the number of rows/columns. So an optimum solution is possible at this stage.
Allot machine Z to job C, Machine Y to job B and Machine X to job A. The total cost
would be: 25 + 20 + 17 = 62. No other assignment would produce a lower cost than the
above.

5.8 TRAVELLING EXECUTIVE/SALESPERSON PROBLEM:

An executive has to visit a number of places. With distance matrix, we can find the
route he has to follow to reduce total distance traveled. The condition is that he should
get back to the place he started first and he should not visit the same place twice. This is
a special case of assignment.

Illustration 7
The following distance matrix among 5 places are given.
To places (figs. In kms)
From A B C D E

107
A 0 44 82 91 65
B 44 0 26 71 36
C 82 26 0 49 80
D 91 71 49 0 59
E 65 36 80 59 0

Find the least cost route if the person has to start from A and get back to A
visiting all places.

Solution
Follow steps 1 and 2 of the assignment model, which diagonal figures A-
A, B-B, C-C, D-D & E-E set to infinity.

With α (infinity) for diagonal moves such moves will go impossible.


Table 1 and 2 gives the resulting figure.
TABLE 1
Row-wise subtraction
A B C D E

A α 38 47 21

B 18 α 0 45 10

C 56 0 α 23 64

D 42 22 0 α 10

E 29 0 44 23 (

TABLE 2
Column wise subtraction
A B C D E
A ( (0) 38 24 11
B (0) ( (0) 21 (0)
C 38 (0) ( (0) 54

108
D 24 22 (0) ( (0)
E (11) 0 44 (0) (

Step 3: The zero cells are our solution cells.

Let him go to B first from A. The zero at A B bracketed. From B he can go to A,


C or E as all has zeros. B to A is no good as he has to visit other places before getting
back to A. So B to C and B to E are the possible solution. B to C is the best choice.
(The reason will be known latter). The zeros at B-A and B-E are cross-marked as these
are not used for the solution. Now he is in C. Then from C to D. Then from D to E.
And finally from E to A. Of course this last move costs 11 kms. Still it is tha available
best, the unused zeros are cross marked. The route is A-B-C-D-E-A. The total distance
is: 44 + 26 + 49 + 59 + 65 = 243 kms.

If instead of B to C, B to E is chosen, then E to D would have been chosen, then D


to C and finally C to A, with extra 38 kms. So this is not good. Here the distance would
be: 270 kms. The last move, with non-zero element, is also bracketed as this is in
solution.

Questions
1. Explain the meaning and uses of transportation and assignment models.
2. Explain the methods of North West corner rule and Vogel’s approximation
method for finding the initial solution.
3. Explain the MODI and stepping stone methods with an example of your own.
4. Explain the mechanism of assignment model.
5. How would you deal with supply-demand inequalities in transportation?
6. Discuss the procedure of assignment model for traveling executive problem.
7. T.C. Mellot trucking company has a contract to move 115 truckloads of sand per
week between three sand-washing plants, W, X and Y, and three destinations, A,
B and C. Cost and volume information is given below. Compute the optimal
transportation cost using the stepping-stone method.

Project Requirement per week, Plant Available per week,


truckloads truckloads
A 45 W 35
B 50 X 40
C 20 Y 40

COST INFORMATION

109
From To Project A To Project B To Project C
Plant W Rs. 5 Rs. 10 Rs. 10
Plant X 20 30 20
Plant Y 5 8 12

8. Sid Lane hauls oranges between Florida groves and citrus packing plants. His
schedule this week calls for 520 boxes with locations and costs as follows:

Grove Available per week Packing plant Requirement per week


A 170 W 130
B 250 X 200
C 100 Y 190

COST INFORMATION
From To plant W To plant X To plant Y
Grove W Rs. 12 Rs. 8 Rs.5
Grove X 11 15 10
Grove Y 2 7 6

9. Jack Evans Owns several trucks used to haul crushed stone to road projects in the
country. The road contractor for whom Jack hauls, N, Teer, has given Jack this
schedule for the next week.

Project Requirement per week Plant Available per week

A 50 W 45
B 75 X 60
C 50 Y 60

Jack figures his cost from the crushing plant to each of the road projects to be
these:

COST INFORMATION:

From To Project A To Project B To Project C


Plant W Rs. 4 Rs. 8 Rs. 3
Plant X 6 7 9
Plant Y 8 2 5

110
10. Coley’s Machine shop has four machines on which to do three jobs. Each job can
be assigned to one and only one machine. The cost of each job on each machine
is given in the following table. What are the job assignment which will minimize
cost.

Machine
Job W X Y Z

A Rs. 18 Rs. 24 Rs. 28 Rs. 32


B 8 13 17 19
C 10 15 19 22

***
LESSON 6
PROBABILITY: CONCEPTS AND USES

One of the statistical tools that is much used in the world of uncertainty by business
managers is probability Jacob Bernoulli, Abraham de Moirre, Rev. Thomas Bayes and
Joseph Lagrange developed probability concepts Pierre Simon and Laplace unified these
concepts. Concept of probability is a part of every-day life.

111
6.1 CONCEPT OF TYPES OF PROBABILITY

Probability means chance. What is the “probability” that the BSE SENSEX will
appreciate by 50 points today over the yesterday’s close? You can replace the term
“probability” by “chance” and represent the above interrogative sentence as follows:
What is the “Chance” that the BSE SENSEX will appreciate 50 points over yesterday’s
close? There is absolutely no difference in meaning between the two sentences.
Probability is thus a science dealing with chance or uncertain outcomes. It deals with
prediction of future uncertain outcomes. Every walk of life is now beset with
uncertainties and that probability science has great value to us to predict the future.
Business managers use probability in a wide variety of situations. There are different
concepts of probability.

6.1.1 Mathematical or Prior or Classical Probability

Mathematically probability is the ratio of number of favourable outcomes to the


number all possible outcomes of an action. Take the action of tossing a coin. What is the
probability of getting head (H)? There are two possible outcome – getting head and
getting tail (T). Therefore number of favourable outcomes (i.e.; getting head, in this case
is one and total number of outcomes is two). So, probability of getting head = No. of
favourable outcome/ Total No. of possible outcomes = ½ = 0.5.

Similar to the above, the probability of getting an even number side up of a dice is =
No. of even number side up of a dice is = No. of even number sides of a dice / total no. of
sides of the dice = 3/6 = 0.5.

Similarly, the probability of drawing an ACE card from a pack of the usual card
game is, number of ACE cards / total number of cards = 4/52 = 0.0769. And probability
of drawing a SPADE = 13/52 = 0.25. And probability of drawing SPADE KING = 1/52
= 0.0192.

The above probability approach cannot be applied to problems other than


experiments with coins, cards and dice. Hence the limitation of priori or mathematical
probability.

6.1.2 Statistical or Empirical Probability or Relative Frequency Approach to


Probability

Statistical probability is the ratio of the number of times a considered


favourable outcome is obtained from an experiment to total number of outcomes
obtained from the experiment to total number of outcomes obtained from the
experiment when the experiment is repeated number of times under stable conditions.

112
What is the probability of the prices of scrips rising when higher divided is
declared? Say of the 300 scrips for which higher (than previous dividend) has been
declared, 240 scrips have reported initial gains. Then the probability = 240 / 300 =
0.8. This approach is known as relative frequency approach.

6.1.3 Subjective or Intuitive Probability

Subjective probability is defined as the probability assigned to an event by an


individual based on whatever evidence is available. Intuition, educated guess or
relative frequency of past occurrences could be assumed as evidence. This approach
is generally adopted when events occur only once or at most a very few times only.

6.2 BASIC TERMS

In probability some concepts are used. We must know each of these concepts.

“Experiment” is a term used, which means the activity performed, like tossing
a coin or playing a dice or drawing a card and so on.

“Outcome” is a term used to depict the results of an experiment. In tossing a


coin head or tail it gets as result.

“Event” is a term which refers to one or more of the possible outcomes


emanating from the experiments.

“Sample space” is a term used which refers to the set of all possible outcome
of an experiment. In a dice game the same space = S = [1, 2, 3, 4, 5 and 6].

“Mutually exclusive” events are those that cannot occur together but only one
at a time. Head or tail and not both can occur when a coin is tossed. So, the events are
mutually exclusive. Sample space of mutually exclusive events is said to contain the list
of mutually exclusive but collectively exhaustive events. ‘Ace’ and ‘King’ are mutually
exclusive events.

“Mutually inclusive” events are those that can occur together. “Ace and
Spade” are mutually inclusive events. Families having CTVs and washing machines –
mutually inclusive.

“Statistical independence” of events means, the outcome of an event has no effect


on the chance of the occurrence of any other event.

“Statistical dependence” of events means, the probability of some event is


dependent upon or affected by the occurrence of some other event.

6.3 MARGINAL, JOINT AND CONDITIONAL PROBABILITIES

113
The concepts of marginal, joint and conditional probabilities are dealt in two
sections.

6.3.1 Statistical Independence


Under conditions of statistical independence the formulas for marginal, joint
and conditional probabilities are attempted below:

i) Marginal Probability or unconditional probability is the simple probability of


the occurrence of an event. It is given by, say P(H) = ½ (1/(1 + 1) = ½ or 0.5
where P(H) is probability of getting head in a coin-toss experiment. The symbol
for marginal probability is P(A) read as probability of event A happening. Each
toss is statistically independence.

ii) Joint Probability is the probability of two or more events happening together or
in succession. It is given by: P(AB) = P(A)*P(B), where P(AB)- is probability
two events A and B happening together or in succession and P(A) and P(B) are
marginal probability of events (A) and (B). in a fair coin tossing experiment
P(H1, H2) – is P of getting heads in successive two losses or getting heads up on
the two coins tossed simultaneously. In a dice game P(2 & 2) = P(2)*P(2) = 1/6 =
1/36.

iii) Conditional Probability is probability of an event happening given that another


event has happened. It is written as P(B/A) and read as probability of event B
happening, given that event A has already happened. Under statistical
independence. P(B/A) = P(B), for no event has influence over other.

6.3.2 Statistical Dependence

Under conditions of statistical dependence marginal and joint probabilities are


some as in statistical independence. But conditional probability differs. Conditional
probability = P(B/A) = P(AB)/P(A), where P(A/B) joint probability and P(A) is marginal
probability. Conditional probability of P(A/B) = P(AB) / P(B).

6.4 RULES FOR ADDITION & MULTIPLICATION


Addition and multiplication rules of probability are dealt now.

6.4.1 Addition Rule

The addition rule for probability is about probability of getting either this or that.
For mutually exclusive events, the addition rule is given by:
P(A or B) = P of (A) or (B) happening = P(A) + P(B)

Note carefully, that addition rule deals with either or situation.

What is the probability that a card drawn is either diamond or spade? P(diamond or
spade) = P(Diamond) + P(Spade) = ¼ + ¼ = ½.

114
For mutually inclusive evens, the addition rule is given by:
P(A or B) = P(A) + P(B) – P(AB), where P(AB) is probability of events A and B
happening together. What is the probability that a card drawn is a spade or king?
P(Spade or king) = P(Spade) + P(king) = ¼ + 1/13 – 1/52 = 16/52 = 4/13.

6.4.2 Multiplication Rule

For multiplication rule of probability is concerned about joint occurrence of events.


For statistically independent events, the joint prob = P(AB) = P(A)P(B). Similarly, P
(ABC) = P(A) P(B) P(C).

For statistically dependent events, P(AB) = P(B/A)


P(A) and P(AB) = P(A/B) – P(B).

Note carefully that multiplication rule is concerned with this and that or both, or
all or joint happening.

Illustration 1
The frequency distribution of brokerage commission earned in a month from a
survey of 300 stock brokers is as follows:

Rs. 0-5000 5000-10000 10000-15000 15000-20000 20000-25000 25000+


Frequency 1525 35 125 70 30
(%) 58.33 11.67 41.67 23.33 10

What is probability that a randomly selected stock broker earns say (i) Rs.5000 to
Rs.10, 000 and (ii) more than Rs. 15, 000.

Solution

The probability that a broker earns between $ 5000 and $ 10000 is 0.0833 or 8.33%
and more than $ 150000 is, 100% minus (P of earning less than $ 5000 + P of earning $
5000 to 10000 + P earning $ 10000 to $ 15000) = 100% - (5% + 8.33% + 11.67%) = 75%
or 0.75. This is a goal example for statistical probability.

Illustration 6.2

Out of every 1000 investors 20 complain non-receipt of share certificates and 18


have their names wrongly spelled. Of these, 5 have both these complaints. What is the P
of any randomly chosen investor has any complaint?

Solution:

115
This is a problem of mutually inclusive type involving addition rule. Let ‘A’ refers
to complaint of non-receipts of share certificate, ‘B’ refers to wrong spelling of name and
therefore ‘AB’ refers to both complaints happening together.

P(A or B) = P(A) + P(B) – P(AB)


= 20/1000 + 18/1000 – 5/1000
= 33/1000 = 0.033 or about 3%.

Illustration 6.3

A research group says, illiquidity of scrips is occurring only one-third as often as


quoting below par. The P of both illiquidity and below par is 0.05. If 80 per cent of
scrips have none of these problems, how low must the illiquidity problem probability be?

Solution

Let ‘A’ denote illiquidity, ‘B’ denote below par and ‘AB’ denote both.
P(A or B) = P(A) + P(B) – P(AB)
P(A or B) = 1 - % of nil problem
= 1 - .8 = .2

So, 0.2 = P(A) + P(B) – P(AB)


So, P(A) + P(B) = 0.2 + 0.05 = 0.25
i.e., P(A) + 3[P(A)] = 0.25 (Since, ‘A’ is a third ‘B’)
So, P(A) = 0.25 / 4 = 0.0625.

Illustration 6.4

An inventor is setting 3 independent criteria for an ideal investment. First it must


be liquid. Second, it must not fall below issue price. Third it must pay commensurate
return. By experience, 85% of investments are found to be liquid. Out of liquid
investments 80% are found to be holding their value. Out of these that hold their value
82% are giving good return. What is the probability that a randomly selected investment
will prove to be ideal.

Solution

This is a case of mutually independent, joint probability problem. The required is


the type P(ABC) given, P(A), P(B) and P(C). So, P(ABC) = P(A)*P(B)*P(C) = (.85)(.8)
(.82) = 0.5576. So there is about 56% chance.

Illustration 6.5

116
A stock broker says out of his clients 63% give “purchase” orders and 32% have
over 3 years association with him. 21% of clients are with 3 years association placing
“purchase” orders. What is the P that a client placing “purchase” order has more than 3
years association?

Solution
This is a case of conditioned probability. Let ‘A’ denote “purchase”, ‘B’ denote
over 3 year standing and ‘AB’ denote those with 3 year placing a “purchase” order.
What is required is P(B/A) = P(BA)/P(A) = 0.21/0.63 = 0.33.

Illustration 6.6

In a survey, the P that a family makes equity investments if its annual income
exceeds Rs.3, 50, 000 is 0.75. Of the surveyed families, 60% have income exceeding
Rs.3, 50, 000 and 52% have equity investments. What is the P that a family has equity
investments? What is the P that a family has equity investments and income over Rs. 3,
50, 000?

Solution

Let ‘A’ denote equity investment and ‘B’ denote income exceeding Rs.3, 50, 000.
P(A) = .52; P(B) = .6 and P(A/B) = 0.75. Required is P(AB). We know that, P(A/B) =
P(BA)/P(B). So, 0.75 = P(BA) / 0.6. So, P(BA) = 0.45.

Illustration 6.7

When a sample of 250 investors, are distributed as to sex and first time or
experienced investors, the following distribution is obtained.

First time Experienced


Male 65 65
Female 44 76
109 141

i) What is P of a randomly selected investor is male?


ii) What is P that the investor is first-time investor given that the person is male?
iii) What is the P that the person is female given that she is an experienced investor?

Solution

The P (male) = Total no. of male/ Total no. of investors


= 130 / 250 = 0.52

117
P(First-time / male) = P(First time and male) / P(Male)
= (65/250) / (130/250) = 65 / 130 = 0.5

P(Female / experienced) = P(Female and experienced) / P(Experienced)


= (76/250) / (141/250) = 76/141 = 0.539

Illustration 6.8

A study of a stock exchange reveals that in the fortnight beginning new-moon, 10


out of 15 days the market grew and in the fortnight beginning full moon on 8 out of 15
days it declined as indicated by the market index. On no day it remained at same level as
at the previous day’s value.
i) Find the P of growth on any randomly selected day, ii) Find the P of growth given
that the day followed full moon day.

Solution
Fortnight Beginning
Full moon New moon Total
Growth 7 10 17
Decline 8 5 13
15 15 30
P(G) = 17 / 30 = 0.567
P(G/F) = P(GF) / P(F)
= (7/30) / (15/30) = 7 / 15 = 0.467

Illustration 6.9

In a sample of 100 people surveyed with debt or equity investments only, 60 were
equity investors and 40 debt investors. Of the 60 equity investors 40 were metro-based
investors and the 20 were non-metro based. Of the 40 debt investors, 10 were metro and
30 non-metro. What is the P that a randomly selected investor is metro, given that he is
an equity investor?

Solution

The required answer is : P(M/E) =P(ME) / P(E)


=(40/100) / (60/100) = 40 / 60 = 0.67
Where M – denotes “metro” and ‘E’ denotes “equity”.

Illustration 6.10

Two cards are drawn from a pack of 52 cards with replacement, i.e., the second
card is drawn after replacing the first card in the pack. Find the probability that, i) both
are king, ii) first is king and second is queen and iii) one is king and other is queen.

118
Solution
i) P(both cards are king) = P(1st king).P(2nd king)
= 4/52 x 4/52 = 1/169

ii) P(first king & second queen) = P(1st king) P(2nd queen)
= 4/52 x 4/52 = 1/169

iii) P(one king and other queen)


= P(first is king & 2nd queen). P(first is queen & 2nd is king)
= (4/52 x 4/52) + (4/52 x 4/52) = 2/169

Illustration 6.11

A pair of dice is thrown 3 times. If getting a doublet is considered a success, find


the probability of i) 3 success, ii) at least 2 successes and iii) at most 2 successes.

Solution
Doubles could be : 1,1; 2,2; 3,3; 4,4; 5,5; 6,6;
Let p = P(1,1; 2,2; 3,3; 4,4; 5,5 or 6,6) = 6/36 = 1/6
q = 1 – p = 1 – 1/6 = 5/6
n = 3 (Number of trials)

i) 3 success: with n = 3, p = 1/6, q = 5/6 and r = 3


P(r = 3) = ncr pr qn-r
= 3c3 (1/6)3 (5/6)(3-3)
= 1 x 1/216 x 1 = 1/216

ii) P(r>2) = P(2 or 3 success here)


We know P(r = 3) = 1/216
P(r = 2) = nc2p2q3-2
= 3 x 1/6 x 1/6 x 5/6 = 15/216
P(r > 2) = 1/216 + 15/216 = 16/216 = 4/54 = 2/27
iii) P(at most 2 success) = P(0) + P(1) + P(2)
= 1 – P(3)
= 1 – 1/216 = 215/216

Illustration 6.12

An urn contains 3 white balls, 4 red balls and 5 black balls. Two balls are drawn:
what is the probability that i) both are red? Ii) both are white? And iii) one red and one
white?

119
Solution
Total balls = 3 + 4 + 5 = 12
i) P(2 red) = 4C2 / 12C2 = [(4 X 3)/2] / [(12 X 11) / 2)] = 1/11
ii) P(2 white) = 3C2 / 12C2 = 3 / (12 X 11 / 2) = 1/22
iii) P(1R & 1W) = [(4C1 X 3C1) / 12C2 = (4X3) / (12X11 / 2) = 2/11

Illustration 6.13

A problem in statistics is given to 3 students. The marginal probability of solving


the problem for each of the students is 0.8, 0.6 and 0.9. What is the probability that the
problem will be solved by one or other?

Solution

The P (Solving) = 1 – p (none solving)


P (None solving) = Joint prob. Of none solving
= (0.2) (0.4) (0.1) = 0.008

P (Solving) = 1 – P(none solving)


= 1 – 0.008 = 0.992

6.5 BAYE’S THEOREM (REVISING PRIOR ESTIMATES OF PROBABILITY)

Baye’s theorem is the basis for conditional probability under statistical


dependence. We know that P(A/B) = P(AB) / P(B) and
P(B/A) = P(AB) / P(A)

Baye’s theorem gives a superior method of evaluating new information and our
earlier estimates (based on limited information only) of the probability revising that
things are in one or other state. Decision making is immensely helped by Baye’s
theorem.

Illustration 6.2.1

In a factory 3 machines A, B & C produce respectively 25%, 35% and 40%. Of


total output of their outputs, 2%, 4% and 5% are defective. A product is randomly
chosen and it is found to be defective. What is the probability that it was produced by i)
A, ii) A or B, iii) C?

Solution

Given P(A) = 0.25, P(B) = 0.35 and P(C) = 0.4


P(D/A) = 0.02, P(D/B) = 0.04 and P(D/C) = 0.05

120
i) P(A/D) = P(A) x P(D/A)
P(A) x P(D/A) + P(B) x P(D/B) + P(D/C)

= 0.25 x 0.02
(0.25 x 0.02) + (0.35 x 0.04) + (0.4 x 0.05)

= 0.005 = 0.005
0.000 + 0.014 – 0.020 0.039

= 0.128

ii) P(C/D) = 1 – P(A/D or B/D)


= 1 – 0.487 = 0.513
(OR)
0.02 / 0.039 = 0.513
Illustration 6.2.2

In a competitive exam, multiple choice questions are used. 4 choices of possible


answers to each question are given, of which 1 answer is correct. An intelligent
candidate knows 90% of answer and below-average candidate knows only 20% of
answers. But guess work also help the candidates gets correct answer. What is the
probability that the intelligent candidate gets it through guess? What is the probability
that a below average candidate got correct answer through guess?

Solution

Let ‘A’ denote knowing the correct answer and ‘α’ not knowing the correct answer.
Let ‘B’ denote getting the correct answer either knowing the answer or through guess.
So, P(B/A) = 1;
P(B/α) = ¼ = 0.25, since one of the 4 answer is correct.

i) Intelligent candidate
P(A) = 0.9 and P(α) = 1 – 0.9 = 0.1

P(α/B) = P(α) x P(B/α)


P(A) x P(B/A) + P(α) x P(B/α)

= 0.1 x 0.25
(0.9 x 1) + (0.1 x 0.25)
= 0.025 = 0.025 = 0.027
0.9 + 0.025 0.925
ii) Below average candidate
P(A) = 0.2 and P(α) = 1 – 0.2 = 0.8

P(α/B) = P(α) x P(B/α)


P(A) x P(B/A) + P(α) x P(B/α)

121
= 0.8 x 0.25
(0.2 x 1) + (0.8 x 0.25)
= 0.2 = 0.2 = 0.5
0.2 + 0.2 0.4

6.6 PROBABILITY AND EXPECTED VALUE

Probability theory helps in computing expected values. Expected return, expected


EPS or DPS, expected risk, expected price, etc. can all be computed using probability
theory. Expected value of anything is the weighted value, the weights being probability
value.

Illustration 6.3.1

Compute expected return on a mutual fund schemes and its risk given the
following:

Return = Ri = 12% 13% 14% 15% 16%


Probability = Pi = 0.1 0.2 0.4 0.2 0.1

Solution

∑(R) = ∑Ri Pi
∑(R) = 12%(0.1) + 13%(0.2) + 14%(0.4) + 15%(0.2) + 16%(0.1)
= 1.2 + 2.6 +5.6 + 3 + 1.6
= 14%

∑ (Risk) = [∑Pi (Ri – E(R)]1/2


= [.1(12 - 14)2 + .2(13 - 14)2 + .4(14 - 14)2 + .2(15 - 14)2 + .1(16 - 14)2]
= [.4 + .2 + 0 + .2 + 4]0.5 = √1.2 = 1.095%

Illustration 6.3.2

The absenteeism on Monday has the following distribution in a Govt. office.

% Absenteeism: 5% 10% 15% 20% 25%


Probability: 40% 30% 20% 6% 4%

Find the expected level of absenteeism.

Solution

Expected absenteeism = ∑PiAi


= 0.4(5%) + 0.3(10%) + 0.2(15%) + 0.06(20%) + 0.04(25%)
= 2% + 3% + 3% + 1.2% + 1%

122
= 10.2% of total staff members.

QUESTIONS

1. Explain the concept and types of probability.

2. Present the uses of probability to investment decision making.

3. Compute expected return and risk of a security whose probability distribution of


returns are:
Ri = 12% 14% 16% 18% 20%
Pi = 0.1 0.15 0.5 0.15 0.1

4. In a survey of 400 investors 50% hold equity investment, 80% hold debt
investments. Find those who hold both debt and equity investments.

5. In an investment game, the choice of investment is made randomly by picking up


applications from a bag which contains 4 forms for share investments, 10 for
debentures investments, 6 for mutual funds and 3 for fixed deposits. Find the
probability of (i) an equity, a debenture, a mutual fund an FD investment, (ii) the
joint probability of (a) two successive mutual fund investment, (b) one equity and
one FD investment.

6. Explain basic concepts of probability theory.

7. A company uses 3 machines, A, B and C. The proportion of products produced


by the machines are 0.3, 0.48 and 0.22. The percent defective for each machine is
1%, 2% and 3% of their individual output. A defective output is drawn from a
day’s production. What is the probability that it was not produced by A?

***

123

You might also like