INDUSTRIAL ENGINEERING

LINEAR PROGRAMMING n n
The general linear programming (LP) problem is: Minimize Ζ = ∑ ∑ ci j xi j
i =1 j =1
Maximize Z = c1x1 + c2x2 + … + cnxn
Subject to: subject to
a11x1 + a12x2 + … + a1nxn ≤ b1 n n
a21x1 + a22x2 + … + a2nxn ≤ b2 ∑ xij − ∑ x ji = bi for each node i∈N
.. j =1 j =1
.
and
am1x1 + am2x2 + … + amnxn ≤ bm
x1, …, xn ≥ 0 0 ≤ xij ≤ uij for each arc (i, j)∈A
An LP problem is frequently reformulated by inserting The constraints on the nodes represent a conservation of
non-negative slack and surplus variables. Although these flow relationship. The first summation represents total flow
variables usually have zero costs (depending on out of node i, and the second summation represents total
the application), they can have non-zero cost coefficients in flow into node i. The net difference generated at node i is
the objective function. A slack variable is used with a "less equal to bi.
than" inequality and transforms it into an equality.
For example, the inequality 5x1 + 3x2 + 2x3 ≤ 5 could be Many models, such as shortest-path, maximal-flow,
changed to 5x1 + 3x2 + 2x3 + s1 = 5 if s1 were chosen as a assignment and transportation models can be reformulated
slack variable. The inequality 3x1 + x2 – 4x3 ≥ 10 might be as minimal-cost network flow models.
transformed into 3x1 + x2 – 4x3 – s2 = 10 by the addition of
STATISTICAL QUALITY CONTROL
the surplus variable s2. Computer printouts of the results of
processing an LP usually include values for all slack and Average and Range Charts
surplus variables, the dual prices, and the reduced costs for n A2 D3 D4
each variable. 2 1.880 0 3.268
3 1.023 0 2.574
Dual Linear Program
4 0.729 0 2.282
Associated with the above linear programming problem is 5 0.577 0 2.114
another problem called the dual linear programming 6 0.483 0 2.004
problem. If we take the previous problem and call it the 7 0.419 0.076 1.924
primal problem, then in matrix form the primal and dual 8 0.373 0.136 1.864
problems are respectively: 9 0.337 0.184 1.816
10 0.308 0.223 1.777
Primal Dual
Xi =
an individual observation
Maximize Z = cx Minimize W = yb n =
the sample size of a group
Subject to: Ax ≤ b Subject to: yA ≥ c k =
the number of groups
x≥0 y≥0 R =
(range) the difference between the largest and
smallest observations in a sample of size n.
It is assumed that if A is a matrix of size [m × n], then y is an
X + X2 +…+ Xn
[1 × m] vector, c is an [1 × n] vector, b is an [m × 1] vector, X= 1
and x is an [n × 1] vector. n
X1 + X 2 + … + X k
Network Optimization X=
k
Assume we have a graph G(N, A) with a finite set of nodes R1 + R2 + … + Rk
N and a finite set of arcs A. Furthermore, let R=
k
N = {1, 2,…, n} The R Chart formulas are:
xij = flow from node i to node j CLR = R
cij = cost per unit flow from i to j UCLR = D4 R
uij = capacity of arc (i, j) LCLR = D3 R
bi = net flow generated at node i The X Chart formulas are:
We wish to minimize the total cost of sending the available CL X = X
supply through the network to satisfy the given demand. The
UCLX = X + A2 R
minimal cost flow model is formulated as shown below:
LCL X = X − A2 R

189
 INDUSTRIAL ENGINEERING (continued)

Standard Deviation Charts PROCESS CAPABILITY

n A3 B3 B4
2 2.659 0 3.267 ⎛ µ − LSL USL − µ ⎞
C pk = min ⎜ , ⎟ , where
3 1.954 0 2.568 ⎝ 3σ 3σ ⎠
4 1.628 0 2.266
5 1.427 0 2.089 µ and σ are the process mean and standard deviation,
6 1.287 0.030 1.970 respectively, and LSL and USL are the lower and upper
7 1.182 0.119 1.882 specification limits, respectively.
8 1.099 0.185 1.815 QUEUEING MODELS
9 1.032 0.239 1.761
Definitions
10 0.975 0.284 1.716
Pn = probability of n units in system,
UCL X = X + A3 S L = expected number of units in the system,
CL X = X Lq = expected number of units in the queue,
LCL X = X − A3 S W = expected waiting time in system,
Wq = expected waiting time in queue,
UCL S = B 4 S
CL S = S
λ = mean arrival rate (constant),
~
LCL S = B3 S λ = effective arrival rate,
Approximations µ = mean service rate (constant),
The following table and equations may be used to generate ρ = server utilization factor, and
initial approximations of the items indicated. s = number of servers.
n c4 d2 d3 Kendall notation for describing a queueing system:
2 0.7979 1.128 0.853 A/B/s/M
3 0.8862 1.693 0.888 A = the arrival process,
4 0.9213 2.059 0.880 B = the service time distribution,
5 0.9400 2.326 0.864 s = the number of servers, and
6 0.9515 2.534 0.848 M = the total number of customers including those in
7 0.9594 2.704 0.833 service.
8 0.9650 2.847 0.820
9 0.9693 2.970 0.808 Fundamental Relationships
10 0.9727 3.078 0.797 L = λW
σˆ = R/d 2 Lq = λWq
W = Wq + 1/µ
σˆ = S/c 4 ρ = λ /(sµ)
σ R = d3σˆ Single Server Models (s = 1)
Poisson Input—Exponential Service Time: M = ∞
σ s = σˆ 1 − c42 , where P0 = 1 – λ/µ = 1 – ρ
σ̂ = an estimate of σ, Pn = (1 – ρ)ρn = P0ρn
σR = an estimate of the standard deviation of the ranges L = ρ/(1 – ρ) = λ/(µ – λ)
of the samples, and Lq = λ2/[µ (µ– λ)]
σS = an estimate of the standard deviation of the standard W = 1/[µ (1 – ρ)] = 1/(µ – λ)
deviations of the samples. Wq = W – 1/µ = λ/[µ (µ – λ)]
Tests for Out of Control Finite queue: M < ∞
~
1. A single point falls outside the (three sigma) control λ = λ (1 − Pn )
limits.
P0 = (1 – ρ)/(1 – ρM+1)
2. Two out of three successive points fall on the same side
of and more than two sigma units from the center line. Pn = [(1 – ρ)/(1 – ρM+1)]ρn
3. Four out of five successive points fall on the same side L = ρ/(1 – ρ) – (M + 1)ρM+1/(1 – ρM+1)
of and more than one sigma unit from the center line. Lq = L – (1 – P0)
4. Eight successive points fall on the same side of the
center line.

190
 INDUSTRIAL ENGINEERING (continued)

Poisson Input—Arbitrary Service Time SIMULATION

2
Variance σ is known. For constant service time, 1. Random Variate Generation
2
σ = 0. The linear congruential method of generating pseudo-
random numbers Ui between 0 and 1 is obtained using
P0 = 1 – ρ Zn = (aZn+C) (mod m) where a, C, m, and Zo are given non-
Lq = (λ2σ2 + ρ2)/[2 (1 – ρ)] negative integers and where Ui = Zi /m. Two integers are
L = ρ + Lq equal (mod m) if their remainders are the same when
Wq = Lq / λ divided by m.
W = Wq + 1/µ 2. Inverse Transform Method
Poisson Input—Erlang Service Times, σ2 = 1/(kµ2) If X is a continuous random variable with cumulative
2
Lq = [(1 + k)/(2k)][(λ )/(µ (µ– λ))] distribution function F(x), and Ui is a random number
between 0 and 1, then the value of Xi corresponding to Ui
= [λ2/(kµ2) + ρ2]/[2(1 – ρ)]
can be calculated by solving Ui = F(xi) for xi. The solution
Wq = [(1 + k)/(2k)]{λ /[µ (µ – λ)]} obtained is xi = F–1(Ui), where F–1 is the inverse function of
W = Wq + 1/µ F(x).
Multiple Server Model (s > 1) F(x)
F(x)
Poisson Input—Exponential Service Times 1

−1 U1 U1
⎡ ⎛ λ⎞n ⎛ λ⎞s ⎤
⎢ s −1 ⎜ ⎟ ⎛ ⎞⎥
⎢ ⎝ µ⎠ ⎝⎜ µ⎠⎟ ⎜ 1 ⎟⎥
P0 = ⎢ ∑ + ⎜ ⎟⎥
⎢ n=0
n! s! ⎜1− λ ⎟ ⎥ U2
⎝ sµ ⎠ ⎥ U2
⎢⎣ ⎦
X X
XX22 0 XX12

⎡ s −1 ( sρ) n ( sρ) s ⎤⎥ Inverse Transform Method for Continuous Random Variables

= 1 ⎢∑ +
⎢⎣ n = 0 n! s!(1 − ρ) ⎥ FORECASTING
⎦
s
Moving Average
⎛ λ⎞ n
P0 ⎜ ⎟ ρ ∑ d t −i
⎝ µ⎠ i =1
Lq = d̂ t = , where
s!(1 − ρ)
2 n
dˆt = forecasted demand for period t,
P0 s s ρs +1
= dt–i = actual demand for ith period preceding t, and
s!(1 − ρ)
2
n = number of time periods to include in the moving
Pn = P0 (λ/µ)n/n! 0≤n≤s average.
n n–s Exponentially Weighted Moving Average
Pn = P0 (λ/µ) /(s! s ) n≥s
Wq = Lq/λ d̂ t = αdt–1 + (1 – α) d̂ t −1 , where

W = Wq + 1/µ d̂ t = forecasted demand for t, and

L = Lq + λ/µ α = smoothing constant, 0≤α≤1

Calculations for P0 and Lq can be time consuming; however,

the following table gives formulas for 1, 2, and 3 servers.
s P0 Lq
2
1 1–ρ ρ /(1 – ρ)
2 (1 – ρ)/(1 + ρ) 2ρ3/(1 – ρ2)
3 2(1 − ρ ) 9ρ 4
2 + 4ρ + 3ρ 2 2 + 2ρ − ρ 2 − 3ρ3

191
 INDUSTRIAL ENGINEERING (continued)

LINEAR REGRESSION Yijkm = average response value for all r2n–2 observations
Least Squares having Xi set at level k, k = 1, 2, and Xj set at level
m, m = 1, 2.
y = â + b̂x , where
Ei = Yi 2 − Yi1
ˆ ,
y - intercept : aˆ = y − bx
Eij =
( ) (
Yij22 − Yij21 − Yij12 − Yij11 )
and slope : bˆ = SS xy /SS xx , 2
n
⎛ n x ⎞⎛ n
⎞ ONE-WAY ANALYSIS OF VARIANCE (ANOVA)
S xy = ∑ xi yi − (1/n ) ⎜ ∑ i ⎟⎜
⎝ i =1 ⎠ ⎝
∑ yi ⎟ ,
⎠ Given independent random samples of size ni from k
i =1 i =1

2
populations, then:
∑ x − (1/n ) ⎛⎜⎝ ∑ x ⎞⎟⎠
n n
k ni
S xx =
( )
2
, 2
i =1
i
i =1
i
∑ ∑ xij − x
i =1 j =1
n = sample size,
k ni
( )
2 k
+ ∑ ni ( xi − x )
2
⎛ n ⎞ = ∑ ∑ xij − x or
y = (1/n ) ⎜ ∑ yi ⎟ , and i =1 j =1 i =1
⎝ i =1 ⎠
SSTotal = SSError + SSTreatments
⎛ n x ⎞.
x = (1/n ) ⎜ ∑ i⎟
⎝ i =1 ⎠
Let T be the grand total of all N=Σini observations and Ti be
the total of the ni observations of the ith sample. See the
Standard Error of Estimate One-Way ANOVA table on page 196.
C = T 2/N
S xx S yy − S xy
2

Se = = MSE , where
2
k ni
S xx ( n − 2 ) SS Total = ∑ ∑ xij2 − C
2
i =1 j =1
n ⎛n ⎞
− (1 n )⎜ ∑ y i ⎟
S yy = ∑
i =1
y i2
⎝ i =1 ⎠
k
SS Treatments = ∑ Ti2 ni − C ( )
i =1
SSError = SSTotal – SSTreatments
Confidence Interval for a
⎛ 1 x2 ⎞
â ± t α 2 ,n − 2 ⎜⎜ + ⎟ MSE
⎟
RANDOMIZED BLOCK DESIGN
⎝ n S xx ⎠ The experimental material is divided into n randomized
blocks. One observation is taken at random for every
Confidence Interval for b
treatment within the same block. The total number of
MSE observations is N=nk. The total value of these observations
b̂ ± t α 2 ,n − 2
S xx is equal to T. The total value of observations for treatment i
is Ti. The total value of observations in block j is Bj.
Sample Correlation Coefficient
S xy
r= C = T 2/N
S xx S yy k n
SS Total = ∑ ∑ xij2 − C
2n FACTORIAL EXPERIMENTS i =1 j =1
Factors: X1, X2, …, Xn n
Levels of each factor: 1, 2 (sometimes these levels are SS Blocks = ∑ ⎛⎜ B 2j k ⎞⎟ − C
represented by the symbols – and +, respectively) j =1⎝ ⎠
r = number of observations for each experimental k
condition (treatment), SS Treatments = ∑ ⎛⎜ Ti2 n ⎞⎟ − C
Ei = estimate of the effect of factor Xi, i = 1, 2, …, n, i =1⎝ ⎠
Eij = estimate of the effect of the interaction between SSError = SSTotal – SSBlocks – SSTreatments
factors Xi and Xj,
n–1
Yik = average response value for all r2 observations See Two-Way ANOVA table on page 196.
having Xi set at level k, k = 1, 2, and

192
 INDUSTRIAL ENGINEERING (continued)
n
ANALYSIS OF VARIANCE FOR 2 FACTORIAL where SSi is the sum of squares due to factor Xi, SSij is the
DESIGNS sum of squares due to the interaction of factors Xi and Xj,
and so on. The total sum of squares is equal to
Main Effects m r T2
Let E be the estimate of the effect of a given factor, let L be SS total = ∑ ∑ Yck2 −
the orthogonal contrast belonging to this effect. It can be c =1k =1 N
proved that where Yck is the k observation taken for the cth experimental
th

L condition, m = 2n, T is the grand total of all observations,

E = n −1 and N = r2n.
2
m RELIABILITY
L = ∑ a (c )Y(c )
c =1
If Pi is the probability that component i is functioning, a
rL2 reliability function R(P1,P2,..,Pn) represents the probability
SS L = , where that a system consisting of n components will work.
2n
m = number of experimental conditions (m = 2n for n For n independent components connected in series,
factors),
n
a(c) = –1 if the factor is set at its low level (level 1) in
R( P1 , P2 , … Pn ) = ∏ Pi
experimental condition c,
i =1
a(c) = +1 if the factor is set at its high level (level 2) in
experimental condition c, For n independent components connected in parallel,
r = number of replications for each experimental
condition n
R( P1 , P2 , … Pn) = 1 − ∏ (1 − Pi )
Y(c ) = average response value for experimental condition c, i =1
and
LEARNING CURVES
SSL = sum of squares associated with the factor.
The time to do the repetition N of a task is given by
Interaction Effects TN = KN s, where
Consider any group of two or more factors. K = constant, and
a(c) = +1 if there is an even number (or zero) of factors in the s = ln (learning rate, as a decimal)/ln 2.
group set at the low level (level 1) in experimental condition If N units are to be produced, the average time per unit is
c = 1, 2,…, m given by
a(c) = –1 if there is an odd number of factors in the group set
at the low level (level 1) in experimental condition Tavg =
K
N (1 + s )
[ ]
(N + 0.5)(1+ s ) − 0.5(1+s )
c = 1, 2…, m
It can be proved that the interaction effect E for the factors INVENTORY MODELS
in the group and the corresponding sum of squares SSL can For instantaneous replenishment (with constant demand
be determined as follows: rate, known holding and ordering costs, and an infinite
stockout cost), the economic order quantity is given by
L
E= 2 AD
n −1
2 EOQ = , where
h
m
L = ∑ a (c ) Υ A = cost to place one order,
( c)
c =1
D = number of units used per year, and
r L2 h = holding cost per unit per year.
SS L =
2n Under the same conditions as above with a finite
replenishment rate, the economic manufacturing quantity is
Sum of Squares of Random Error
given by
The sum of the squares due to the random error can be
computed as 2 AD
EMQ = , where
h(1 − D R )
SSerror = SStotal – ΣiSSi – ΣiΣjSSij – … – SS12…n
R = the replenishment rate.

193
 INDUSTRIAL ENGINEERING (continued)

ERGONOMICS
NIOSH Formula PERMISSIBLE NOISE EXPOSURES *
Recommended Weight Limit (U.S. Customary Units)
= 51(10/H)(1 – 0.0075|V – 30|)(0.82 + 1.8/D)(1 – 0.0032A) Duration per day, hours Sound level, dBA
where 8 90
H = horizontal distance of the hand from the midpoint
of the line joining the inner ankle bones to a point 6 92
projected on the floor directly below the load
center, 4 95
V = vertical distance of the hands from the floor, 3 97
D = vertical travel distance of the hands between the
origin and destination of the lift, and 2 100
A = asymmetric angle, in degrees. 1 1/2 102
The NIOSH formula as stated here assumes that (1) lifting
frequency is no greater than one lift every 5 minutes; (2) the 1 105
person can get a good grip on the object being lifted.
1/2 110
Biomechanics of the Human Body
1/4 or less 115
* Exposure to impulsive or impact noise should not exceed
140 dB peak sound pressure level.

FACILITY PLANNING
Equipment Requirements
Basic Equations Pij = desired production rate for product i on machine j,
Hx + Fx = 0 measured in pieces per production period,
Hy + Fy = 0 Tij = production time for product i on machine j,
Hz + Fz = 0 measured in hours per piece,
THxz + TFxz = 0 Cij = number of hours in the production period available
for the production of product i on machine j,
THyz + TFyz = 0
Mj = number of machines of type j required per
THxy + TFxy = 0 production period, and
n = number of products.
The coefficient of friction µ and the angle α at which the Therefore, Mj can be expressed as
floor is inclined determine the equations at the foot. n Pij Tij
Fx = µFz Mj =∑
i =1 Cij
With the slope angle α
Fx = µFzcos α
Of course, when motion must be considered, dynamic
conditions come into play according to Newton's Second
Law. Force transmitted with the hands is counteracted at the
foot. Further, the body must also react with internal forces at
all points between the hand and the foot.

194
 INDUSTRIAL ENGINEERING (continued)

People Requirements Material Handling

n Pij Tij Distances between two points (x1, y1) and (x1, y1) under
Aj = ∑ , where different metrics:
i =1 Cij
Euclidean:
Aj = number of crews required for assembly operation j,
Pij = desired production rate for product i and assembly D= (x1 − x2 )2 + ( y1 − y 2 )2
operation j (pieces per day),
Rectilinear (or Manhattan):
Tij = standard time to perform operation j on product i
(minutes per piece), D = ⏐x1 – x2⏐ + ⏐y1 – y2⏐
Cij = number of minutes available per day for assembly Chebyshev (simultaneous x and y movement):
operation j on product i, and
D = max(⏐x1 – x2⏐ , ⏐y1 – y2⏐)
n = number of products.
Line Balancing
Standard Time Determination
⎛ ⎞
ST = NT × AF Nmin = ⎜ OR × ∑ ti OT ⎟
⎝ i ⎠
where
= Theoretical minimum number of stations
NT = normal time, and
Idle Time/Station = CT – ST
AF = allowance factor.
Idle Time/Cycle = Σ (CT – ST)
Case 1: Allowances are based on the job time.
AFjob = 1 + Ajob Idle Time Cycle
Percent Idle Time = × 100 , where
N actual × CT
Ajob = allowance fraction (percentage) based on job time.
CT = cycle time (time between units),
Case 2: Allowances are based on workday.
OT = operating time/period,
AFtime = 1/(1 – Aday)
OR = output rate/period,
Aday = allowance fraction (percentage) based on workday.
ST = station time (time to complete task at each station),
ti = individual task times, and
Plant Location
N = number of stations.
The following is one formulation of a discrete plant location
problem. Job Sequencing
Minimize Two Work Centers—Johnson's Rule
m n n 1. Select the job with the shortest time, from the list of
z = ∑ ∑ cij yij + ∑ f j x j jobs, and its time at each work center.
i =1 j =1 j =1
2. If the shortest job time is the time at the first work
subject to
center, schedule it first, otherwise schedule it last. Break
m
ties arbitrarily.
∑ y ij ≤ mx j , j = 1,… , n
i =1 3. Eliminate that job from consideration.
n 4. Repeat 1, 2, and 3 until all jobs have been scheduled.
∑ y ij = 1, i = 1,… , m
j =1 CRITICAL PATH METHOD (CPM)
yij ≥ 0, for all i, j dij = duration of activity (i, j),
xj = (0, 1), for all j, where CP = critical path (longest path),
m = number of customers, T = duration of project, and
n = number of possible plant sites, T = ∑ d ij
yij = fraction or portion of the demand of customer i (i , j )∈CP
which is satisfied by a plant located at site j; i = 1,
…, m; j = 1, …, n,
xj = 1, if a plant is located at site j,
xj = 0, otherwise,
cij = cost of supplying the entire demand of customer i
from a plant located at site j, and
fj = fixed cost resulting from locating a plant at site j.

195
 INDUSTRIAL ENGINEERING (continued)

PERT TAYLOR TOOL LIFE FORMULA

(aij, bij, cij) = (optimistic, most likely, pessimistic) VT n = C, where
durations for activity (i, j),
V = speed in surface feet per minute,
µij = mean duration of activity (i, j),
T = time before the tool reaches a certain percentage of
σij = standard deviation of the duration of activity (i, j),
possible wear, and
µ = project mean duration, and
C, n = constants that depend on the material and on the
σ = standard deviation of project duration.
tool.
aij + 4bij + cij
µ ij = WORK SAMPLING FORMULAS
6
cij − aij p(1 − p ) 1− p
σ ij = D = Zα 2 and R = Zα 2 , where
6 n pn
µ = ∑ µ ij
(i , j )∈CP p = proportion of observed time in an activity,
2
σ = ∑ σ ij2 D = absolute error,
(i , j )∈CP
R = relative error (R = D/p), and
n = sample size.

ONE-WAY ANOVA TABLE

Degrees of Sum of
Source of Variation Mean Square F
Freedom Squares
SSTreatments MST
Between Treatments k–1 SSTreatments MST =
k −1 MSE
SS Error
Error N–k SSError MSE =
N −k

Total N–1 SSTotal

TWO-WAY ANOVA TABLE

Degrees of Sum of
Source of Variation Mean Square F
Freedom Squares
SSTreatments MST
Between Treatments k–1 SSTreatments MST =
k −1 MSE
SS Blocks MSB
Between Blocks n–1 SSBlocks MSB =
n −1 MSE
SS Error
Error (k–1)(n–1) SSError MSE =
(k − 1 )(n − 1 )

Total N–1 SSTotal

196
 INDUSTRIAL ENGINEERING (continued)

PROBABILITY AND DENSITY FUNCTIONS: MEANS AND VARIANCES

Variable Equation Mean Variance

Binomial ⎛ n⎞ n!
⎜ ⎟=
Coefficient ⎝ x ⎠ x!(n − x )!

⎛ n⎞
b(x; n , p ) = ⎜ ⎟ p x (1 − p )
n− x
Binomial np np(1 – p)
⎝ x⎠

⎛N − r⎞
⎜ ⎟
Hyper ⎛r ⎞ ⎝n − x ⎠ nr r (N − r )n(N − n )
h( x ; n , r , N ) = ⎜ ⎟
Geometric ⎝ x⎠ ⎛ N ⎞ N N 2 (N − 1)
⎜ ⎟
⎝n ⎠

λx e − λ
Poisson f (x; λ ) = λ λ
x!

Geometric g(x; p) = p (1 – p)x–1 1/p (1 – p)/p2

Negative ⎛ y + r − 1⎞ r
f ( y; r , p ) = ⎜ ⎟ p (1 − p )
y
r/p r (1 – p)/p2
Binomial ⎝ r − 1 ⎠
n!
Multinomial f (x1 ,… xk ) = p1x1 … pkxk npi npi (1 – pi)
x1! ,… , xk !

Uniform f(x) = 1/(b – a) (a + b)/2 (b – a)2/12

x α −1e − x β
Gamma f (x ) = ; α > 0, β > 0 αβ αβ2
β α Γ(α )

1 −x β
Exponential f (x ) = e β β2
β
α α −1 − x α ⎡ ⎛ α + 1⎞ 2 ⎛ α + 1 ⎞⎤
Weibull f (x ) = x e β
β1 α Γ[(α + 1) α ] β 2 α ⎢ Γ⎜ ⎟−Γ ⎜ ⎟⎥
β ⎣ ⎝ α ⎠ ⎝ α ⎠⎦
2
1 ⎛ x −µ ⎞
1 − ⎜ ⎟
Normal f (x ) = e 2⎝ σ ⎠
µ σ2
σ 2π

⎧ 2(x − a )
⎪⎪ (b − a )(m − a ) if a ≤ x ≤ m
Triangular f (x ) = ⎨
a+b+m a 2 + b 2 + m 2 − ab − am − bm
2(b − x ) 3
⎪ if m < x ≤ b 18
⎪⎩ (b − a )(b − m )

197
 INDUSTRIAL ENGINEERING (continued)

HYPOTHESIS TESTING
Table A. Tests on means of normal distribution—variance known.
Hypothesis Test Statistic Criteria for Rejection
H0: µ = µ0
H1: µ ≠ µ0 |Z0| > Zα/2

H0: µ = µ0 X − µ0
Z0 ≡ Z0 < – Z α
H0: µ < µ0 σ n
H0: µ = µ0
H1: µ > µ0 Z0 > Z α

H0: µ1 – µ2 = γ
H1: µ1 – µ2 ≠ γ |Z0| > Zα/2

X1 − X 2 − γ
H0: µ1 – µ2 = γ Z0 ≡
H1: µ1 – µ2 < γ σ12 σ 22 Z0 <– Zα
+
n1 n2
H0: µ1 – µ2 = γ
H1: µ1 – µ2 > γ Z0 > Z α

Table B. Tests on means of normal distribution—variance unknown.

Hypothesis Test Statistic Criteria for Rejection
H0: µ = µ0
H1: µ ≠ µ0 |t0| > tα/2, n – 1

H0: µ = µ0 X − µ0
t0 = t0 < –t α, n – 1
H1: µ < µ0 S n
H0: µ = µ0
t0 > tα, n – 1
H1: µ > µ0
X1 − X 2 − γ
t0 =
H0: µ1 – µ2 = γ 1 1
Sp + |t0| > t α/2, v
H1: µ1 – µ2 ≠ γ n1 n2
v = n1 + n2 − 2

X1 − X 2 − γ
H0: µ1 – µ2 = γ t0 =
S12 S22 t0 < –t α, v
H1: µ1 – µ2 < γ +
n1 n2
2
⎛ S12 S22 ⎞
⎜n +n ⎟
H0: µ1 – µ2 = γ ⎝ 1 2⎠
ν= t0 > tα, v
(S n ) + (S )
2 2
H1: µ1 – µ2 > γ 2 2
n2
1 1 2
n1 − 1 n2 − 1
In Table B, Sp2 = [(n1 – 1)S12 2
+ (n2 – 1)S2 ]/v

198
 INDUSTRIAL ENGINEERING (continued)

Table C. Tests on variances of normal distribution with unknown mean.

Hypothesis Test Statistic Criteria for Rejection
H0: σ2 = σ02 χ 02 > χ 2 α 2 ,n −1 or
H1: σ2 ≠ σ02 χ 02 < χ 21− α 2 ,n −1
H0: σ2 = σ02
χ 02 =
(n − 1)S 2
H1: σ2 < σ02 χ 02 < χ 2 1–α/2, n – 1
σ02

H0: σ2 = σ02
H1: σ2 > σ02 χ 02 > χ 2 α, n – 1

H0: σ12 = σ22 S12 F0 > Fα 2 , n1 −1, n2 −1

F0 =
H1: σ12 ≠ σ22 S 22 F0 < F1−α 2 , n1 −1, n2 −1

H0: σ12 = σ22 S 22

H1: σ12 < σ22 F0 = F0 > Fα , n2 −1, n1 −1
S12

H0: σ12 = σ22 S12

F0 = F0 > Fα , n1 −1, n2 −1
H1: σ12 > σ22 S 22

ANTHROPOMETRIC MEASUREMENTS

SITTING HEIGHT SITTING HEIGHT (NORMAL)

(ERECT)

THIGH CLEARANCE
HEIGHT ELBOW-TO-ELBOW
BREADTH
KNEE HEIGHT
ELBOW REST SEAT BREADTH
HEIGHT

BUTTOCK BUTTOCK
POPLITEAL KNEE LENGTH
LENGTH

POPLITEAL
HEIGHT (AFTER SANDERS AND McCORMICK,
HUMAN FACTORS IN DESIGN, McGRAW HILL, 1987)

199
 INDUSTRIAL ENGINEERING (continued)

ERGONOMICS

U.S. Civilian Body Dimensions, Female/Male, for Ages 20 to 60 Years

(Centimeters)
(See Anthropometric Percentiles
Measurements Figure) 5th 50th 95th Std. Dev.
HEIGHTS
Stature (height) 149.5 / 161.8 160.5 / 173.6 171.3 / 184.4 6.6 / 6.9
Eye height 138.3 / 151.1 148.9 / 162.4 159.3 / 172.7 6.4 / 6.6
Shoulder (acromion) height 121.1 / 132.3 131.1 / 142.8 141.9 / 152.4 6.1 / 6.1
Elbow height 93.6 / 100.0 101.2 / 109.9 108.8 / 119.0 4.6 / 5.8
Knuckle height 64.3 / 69.8 70.2 / 75.4 75.9 / 80.4 3.5 / 3.2
Height, sitting 78.6 / 84.2 85.0 / 90.6 90.7 / 96.7 3.5 / 3.7
Eye height, sitting 67.5 / 72.6 73.3 / 78.6 78.5 / 84.4 3.3 / 3.6
Shoulder height, sitting 49.2 / 52.7 55.7 / 59.4 61.7 / 65.8 3.8 / 4.0
Elbow rest height, sitting 18.1 / 19.0 23.3 / 24.3 28.1 / 29.4 2.9 / 3.0
Knee height, sitting 45.2 / 49.3 49.8 / 54.3 54.5 / 59.3 2.7 / 2.9
Popliteal height, sitting 35.5 / 39.2 39.8 / 44.2 44.3 / 48.8 2.6 / 2.8
Thigh clearance height 10.6 / 11.4 13.7 / 14.4 17.5 / 17.7 1.8 / 1.7
DEPTHS
Chest depth 21.4 / 21.4 24.2 / 24.2 29.7 / 27.6 2.5 / 1.9
Elbow-fingertip distance 38.5 / 44.1 42.1 / 47.9 46.0 / 51.4 2.2 / 2.2
Buttock-knee length, sitting 51.8 / 54.0 56.9 / 59.4 62.5 / 64.2 3.1 / 3.0
Buttock-popliteal length, sitting 43.0 / 44.2 48.1 / 49.5 53.5 / 54.8 3.1 / 3.0
Forward reach, functional 64.0 / 76.3 71.0 / 82.5 79.0 / 88.3 4.5 / 5.0
BREADTHS
Elbow-to-elbow breadth 31.5 / 35.0 38.4 / 41.7 49.1 / 50.6 5.4 / 4.6
Hip breadth, sitting 31.2 / 30.8 36.4 / 35.4 43.7 / 40.6 3.7 / 2.8
HEAD DIMENSIONS
Head breadth 13.6 / 14.4 14.54 / 15.42 15.5 / 16.4 0.57 / 0.59
Head circumference 52.3 / 53.8 54.9 / 56.8 57.7 / 59.3 1.63 / 1.68
Interpupillary distance 5.1 / 5.5 5.83 / 6.20 6.5 / 6.8 0.4 / 0.39
HAND DIMENSIONS
Hand length 16.4 / 17.6 17.95 / 19.05 19.8 / 20.6 1.04 / 0.93
Breadth, metacarpal 7.0 / 8.2 7.66 / 8.88 8.4 / 9.8 0.41 / 0.47
Circumference, metacarpal 16.9 / 19.9 18.36 / 21.55 19.9 / 23.5 0.89 / 1.09
Thickness, metacarpal III 2.5 / 2.4 2.77 / 2.76 3.1 / 3.1 0.18 / 0.21
Digit 1
Breadth, interphalangeal 1.7 / 2.1 1.98 / 2.29 2.1 / 2.5 0.12 / 0.13
Crotch-tip length 4.7 / 5.1 5.36 / 5.88 6.1 / 6.6 0.44 / 0.45
Digit 2
Breadth, distal joint 1.4 / 1.7 1.55 / 1.85 1.7 / 2.0 0.10 / 0.12
Crotch-tip length 6.1 / 6.8 6.88 / 7.52 7.8 / 8.2 0.52 / 0.46
Digit 3
Breadth, distal joint 1.4 / 1.7 1.53 / 1.85 1.7 / 2.0 0.09 / 0.12
Crotch-tip length 7.0 / 7.8 7.77 / 8.53 8.7 / 9.5 0.51 / 0.51
Digit 4
Breadth, distal joint 1.3 / 1.6 1.42 / 1.70 1.6 / 1.9 0.09 / 0.11
Crotch-tip length 6.5 / 7.4 7.29 / 7.99 8.2 / 8.9 0.53 / 0.47
Digit 5
Breadth, distal joint 1.2 / 1.4 1.32 / 1.57 1.5 / 1.8 0.09/0.12
Crotch-tip length 4.8 / 5.4 5.44 / 6.08 6.2 / 6.99 0.44/0.47
FOOT DIMENSIONS
Foot length 22.3 / 24.8 24.1 / 26.9 26.2 / 29.0 1.19 / 1.28
Foot breadth 8.1 / 9.0 8.84 / 9.79 9.7 / 10.7 0.50 / 0.53
Lateral malleolus height 5.8 / 6.2 6.78 / 7.03 7.8 / 8.0 0.59 / 0.54

Weight (kg) 46.2 / 56.2 61.1 / 74.0 89.9 / 97.1 13.8 / 12.6

200
 INDUSTRIAL ENGINEERING (continued)

ERGONOMICS—HEARING
The average shifts with age of the threshold of hearing for pure tones of persons with "normal" hearing, using a 25-year-old
group as a reference group.

Equivalent sound-level contours used in determining the A-weighted sound level on the basis of an octave-band analysis. The
curve at the point of the highest penetration of the noise spectrum reflects the A-weighted sound level.

201
 INDUSTRIAL ENGINEERING (continued)

Estimated average trend curves for net hearing loss at 1,000, 2,000, and 4,000 Hz after continuous exposure to steady noise.
Data are corrected for age, but not for temporary threshold shift. Dotted portions of curves represent extrapolation from
available data.

1 2 4 8 10 20 40

Tentative upper limit of effective temperature (ET) for unimpaired mental performance as related to exposure time; data are
based on an analysis of 15 studies. Comparative curves of tolerable and marginal physiological limits are also given.

ATMOSPHERE CONDITIONS

120 48.9

110 43.4
EFFECTIVE TEMPERATURE, °C
EFFECTIVE TEMPERATURE, °F

MARGINAL PHYSIOLOGICAL LIMIT

100 37.8

TOLERABLE PHYSIOLOGICAL LIMIT

90 32.2

UPPER LIMIT FOR

UNIMPAIRED
80 MENTAL PERFORMANCE 26.7

0 60 120 180 240 300 360

EXPOSURE TIME MINUTES
I d ti l E i i

202