Applied Mathematical Modelling 37 (2013) 96989706

Contents lists available at SciVerse ScienceDirect

Applied Mathematical Modelling

journal homepage: www.elsevier.com/locate/apm

An Economic order quantity model for Items with

Three-parameter Weibull distribution Deterioration,
Ramp-type Demand and Shortages
S.S. Sanni , W.I.E. Chukwu
Department of Statistics, Faculty of Physical Sciences, University of Nigeria, Nsukka, Enugu, Nigeria

a r t i c l e

i n f o

Article history:
Received 30 May 2012
Received in revised form 16 April 2013
Accepted 21 May 2013
Available online 5 June 2013
Keywords:
Inventory model
Ramp type demand
Differential equation
Optimal policy
Instantaneous rate function

a b s t r a c t
In this paper, we develop an economic order quantity inventory model for items with
three-parameter Weibull distribution deterioration and ramp-type demand. Shortages
are allowed in the inventory system and are completely backlogged. The demand rate is
deterministic and varies with time up to a certain point and eventually stabilized and
becomes constant. The instantaneous rate of deterioration is an increasing function of time.
We provide simple analytical tractable procedures for deriving the model and give numerical examples to illustrate the solution procedure. Our adoption of ramp-type demand
reects a real market demand for newly launched product.
2013 Elsevier Inc. All rights reserved.

1. Introduction
Inventory control policy is concerned basically with two decisions ; How much to order (produce or purchase) to replenish the inventory of an item and When to order so as to minimize the total cost [1]. Several inventory models have been
developed to answer the above questions. The economic-order-quantity model, hereafter EOQ, was originally developed by
Ford W. Harris in 1913, but R. H. Wilson a consultant who applied it extensively, is given credit for his in-depth analysis [2].
The basic EOQ model assumes a constant demand rate and an innite planning horizon, there is no deterioration of inventory
and replenishment is instantaneous i.e. the lead time is zero. These assumptions restrict the applicability of the classical EOQ
model. In order to make the basic EOQ model more realistic, many researchers have extended Wilsons EOQ model by considering time/price varying demand pattern and deterioration rate. It ispertinent to note that the depletion of any inventory
is due mainly to demand and partly to deterioration of the item.
In what follows we give denition of terms central to the paper. In particular, we dene deterioration, ramp-type function
and Weibull function.
Denition 1.1. Deterioration is dened as decay, damage, change or spoilage that prevents items from being used for its original
purpose. Some examples of items that deteriorate are fashion goods, foods, mobile phones, chemicals, automobiles, drugs, etc.

Denition 1.2. A random variable Y (e.g. the time to deterioration of an item) is said to have a Weibull distribution if its
b
density is given, for some parameters a > 0; b > 0; and c, by f y aby  cb1 eatc ; y > 0. The parameters a; b, and c
Corresponding author. Tel.: +234 8035813754.
E-mail addresses: sanni_smaila@yahoo.com (S.S. Sanni), chukwuagusi@yahoo.com (W.I.E. Chukwu).
0307-904X/$ - see front matter 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.apm.2013.05.017

S.S. Sanni, W.I.E. Chukwu / Applied Mathematical Modelling 37 (2013) 96989706

9699

are the the scale, shape and location parameters of the distribution, respectively. For time to deterioration data, the scale
parameter, a represents the characteristic rate of deterioration of items, the shape parameter, b is a measure of the spread
of time-to-deterioration and c is the minimum time such that y P c. A negative c may indicate that deterioration has
occurred prior to the beginning of the inventory, namely during production, in transit, or prior to actual storage while
c 0 reduces the density to the case of two-parameter Weibull distribution. The scale and the location parameters have
the same unit as T (i.e. cycle length in time) while the shape parameter is dimensionless. Weibull model is widely used today
for failure and survival analysis.

hx; x < l
Denition 1.3. Let R be the set of real numbers. A function g : R ! R dened by gx
is called a ramp

hl; x P l
0; x < l
and a0; l > 0 is a ramp-type function.
function. In particular, gx ax  x  lHx  l where Hx  l
1; x P l
A variety of ramp-type functions have evolved from the above denition.
Inventory models for deteriorating items have been widely studied by researchers. Inventory problem for deteriorating
items was rst studied by Whitin [3], he considered fashion items decaying at the end of the planning horizon. Wagner
and Whitin[4] developed a dynamic version of the classical EOQ model. Thereafter, Ghare and Schrader[5] developed a model for exponentially deteriorating inventory. They proposed, explicitly for the rst time, the differential equation governing
the variation in the inventory system ; dIt=dt hIt Dt. Donaldson[6] provided a somehow complicated analytical
solution procedure for the basic inventory policy for the case of positive linear trend in demand. Silver[7] formulated a heuristic for deteriorating inventory model with time-dependent linear demand. Deb and Chuadhuri[8] extended the deteriorating inventory model with linear demand by incorporating shortages in the inventory. Covert and Philip [9] presented
an inventory model where the time to deterioration is described with two-parameter Weibull distribution. Philip[10] generalized the model in [9] by considering a three-parameter Weibull distribution deterioration, no shortages and a constant
demand. Chakrabarty et al. [11] proposed an EOQ model with three-parameter Weibull distribution deterioration, shortages
and linear demand rate and obtained innite series representation for the initial inventory level and the average total cost
equation. Ghosh and Chaudhuri [12] developed an inventory model for two-parameter Weibull deteriorating items, with
shortages and quadratic demand rate and gave innite series representation for the initial inventory level and the total average variable cost equation. Sanni [13] developed an inventory model with three-parameter Weibull deterioration, shortages
and quadratic demand rate. He derived explicit equations for the initial inventory level and the average total cost by approximating exponential functions by the rst two terms of the Tailor series. An order-level inventory model for deteriorating
items with ramp type demand rate was discussed by Mandal and Pal [14]. They considered an EOQ model for items with
constant rate of deterioration, ramp-type demand rate and no shortage allowed in the system and obtained an approximate
solution for the EOQ. This work was extended by Wu and Ouyang [15] by considering two types of shortages in the model:
model that starts with stock and model that starts with shortages and obtained optimal replenishment policy for the different cases. This model was further generalized by Samanta and Bhowmick [16] by taking two parameter Weibull distribution
to represent the time to deterioration and allowed shortages in the inventory.They studied two cases; where the inventory
starts with shortages and the case where the system starts without shortages and derived the EOQ for the respective systems. For some literature on deteriorating inventory models, see [12,17,18].
In this paper we consider the problem of nding optimal replenishment policy for an inventory system which holds items
with three parameter Weibull Distribution deterioration, ramp type demand rate and shortages are allowed and completely
backlogged. The research focus of this paper is to develop a mathematical model for the system, provide an optimal replenishment policy for the model and establish the necessary and sufcient conditions for the optimal policy. The Weibull distribution is suitable for items whose rate of deterioration increases with time and the location parameter c, in the threeparameter Weibull distribution, is used here to depict the item shelf-life; an important feature of most deteriorating items.
The ramp type demand rate describes the demand of products, such as fashion goods, electronics, automobiles, etc, for which
the demand increases as they are launched into the market and after some time the demand stabilizes and becomes
constant.
2. Assumptions and notation
The mathematical model in this work is developed on the basis of the following assumptions and notation.
Notation
C1
C2
C3
C4
Dt
T
Io

:
:
:
:
:
:
:

inventory holding cost per unit per unit time.

shortage cost per unit per unit time.
ordering cost per order.
unit cost.
demand rate at any time t P 0.
cycle length.
size of the initial inventory.

9700

S.S. Sanni, W.I.E. Chukwu / Applied Mathematical Modelling 37 (2013) 96989706

ht abt  cb1 : instantaneous rate function for a three-parameter Weibull distribution; where a is the scale parameter,
0 < a  1; b is the shape parameter, b > 0 and c is the location parameter, c > 0. Also, t is time to deterioration, t P c
t1 : time during which there is no shortage.
T  : optimal value of T.
Io : optimal value of Io .
t1 : optimal value of t 1 .
Assumptions
(i) The inventory system under consideration deals with single item. This assumption ensures that a single item is isolated from other items and thus preventing item interdependencies.
(ii) The time horizon is innite and a typical planning schedule of cycle of length T is considered.
(iii) The demand rate is a ramp type function of time, i.e. Dt at  t  lHt  l where a and l are constants such

1; t P l
that l > 0, and Ht  l is a Heaviside unit function of time dened as follows: Ht  l
. Also, a stands
0; t < l
for the initial demand rate and l is a xed point in time. The implication of the ramp-type demand rate is that demand
varies linearly with time up to some time l and then stays constant.
(iv) Shortages in the inventory are allowed and completely backlogged so that at inventory level zero all arrived demand is
permitted.
(v) Replenishment is instantaneous and lead time is zero.The assumption of zero lead time is made so that the period of
shortage is not affected.
(vi) Deteriorated unit is not repaired or replaced during a given cycle.
(vii) The holding cost, ordering cost, shortage cost and unit cost remain constant over time.
(viii) There are no quantity discounts.
(ix) The distribution of the time to deterioration of the items follows three-parameter Weibull distribution, i.e
b
f t abt  cb1 eatc ; t > 0. The instantaneous rate function is ht abt  cb1 . The implication of the threeparameter Weibull instantaneous rate is that the impact of the already deteriorated items that are received into
the inventory as well as those items that may start deteriorating in future are accounted for. The graph of three-paprameter Weibull distribution- time relationship is given in Fig. 1 below.
The essence of the assumptions is to make the complexity of the inventory system malleable to mathematical modelling.
The assumptions are selected to give accurate approximation of real life inventory system for newly introduced product.
Remark 1. The assumption of zero time is crucial to the mathematics of the model. If a positive lead time, L is assumed three
possible situations may arise and the situations may be presented as follows:
Let lL be a prescribed interval between orders, then
situation 1. the prescribed time lL starts with inventory level say M 1 and ends with inventory level say M 2 , i.e.
0 < lL < l.
situation 2. the prescribe time lL starts with inventory level say M 1 and ends with a shortage say M 2 , i.e. l < lL < t 1 .
situation 3. the prescribed time lL starts with a shortage says M 1 and ends with a shortage say M 2 , i.e. t1 < lL < T.
These three situations affect the shortage period and thus complicate the mathematics of the model if a positive lead time
L > 0 is assumed.
Motivated by the on-going research on inventory models for deteriorating items, it is our purpose in this paper to provide
optimal inventory policy for the EOQ model with three-parameter Weibull distribution deterioration, ramp-type demand
and shortages. And also establish the necessary and sufcient conditions for this optimal policy. Our model is a variant of
Samanta and Bhowmick [16] and our results extend some results of [16] and many other recently known results in literature.

Rate of deterioration

Increasing rate ( 2 > > 1)

<0
=0
>0

Decreasing rate, < 0

( < 1)

Time
Fig. 1. Rate of deterioration-time relationship for three-parameter Weibull distribution.

9701

S.S. Sanni, W.I.E. Chukwu / Applied Mathematical Modelling 37 (2013) 96989706

3. Mathematical model and analysis

At the start of the cycle, the inventory level reaches its maximum I0 units of item at time t 0. During the time interval
0; t1 , the inventory depletes due mainly to demand and partly to deterioration. At time t l < t1 , the inventory level depletes to S units and at t 1 , the inventory level is zero and all the demand hereafter (i.e. T  t1 is completely backlogged. The
total number of backordered items is replaced by the next replenishment. The demand varies with time up to a certain time
and become constant. The deterioration rate is described by an increasing function of time ht abt  cb1 . A Graphical
representation of the considered inventory system is given below (Fig. 2).
The changes in the inventory at any time t are governed by the differential equations:

dIt
Itabt  cb1 at;
dt

0 6 t 6 l;

3:1

dIt
Itabt  cb1 al;
dt

l 6 t 6 t1 ;

3:2

dIt
al;
dt

t1 6 t 6 T;

3:3

with boundary conditions I0 I0 ; Il S; It1 0 and IT 0. Using the assumptions and the boundary conditions in Eq.
3.1, Eq. 3.2 and Eq. 3.3, we have

"

"
#
#
t 2 att  cb1 at  cb2  cb2
I0 expfacb g expfat  cb g;


b 1b 2
2
b1
" "
##
b1
at1  c  t  cb1
expfat  cb g; l 6 t 6 t 1 ;
It al t1  t
b 1

It a

It alt  t 1 ;

0 6 t 6 l;

t 1 6 t 6 T:

3:4
3:5
3:6

Applying the condition Il S in Eqs. 3.4 and 3.5, we get

"

"

S a
"

l2
2

all  cb1
b1

"

al  cb2  cb2

#
b

I0 expfac g expfal  cb g;

b 1b 2

3:7

##

S al t1  l

at1  cb1  l  cb1

b 1

expfal  cb g:

3:8

Eliminate S by equating Eq. 3.7 and 3.8, we obtain the initial inventory level

"

#

l alt1  cb1 al  cb2  cb2
I0 a l t 1 


expfacb g
b 1b 2
2
b1
"
#


l alt1  cb1 al  cb2  cb2
l
:
 acb l t 1 


a l t1 
b 1b 2
2
b1
2

3:9

Inventory level

Depletion curve with detarioration

I0
Without detarioration

S
I

t1
t=

( I I0 )
T
Fig. 2. Inventory system for deteriorating items.

Time

9702

S.S. Sanni, W.I.E. Chukwu / Applied Mathematical Modelling 37 (2013) 96989706

The total inventory cost per unit time consists of the the following components:
Deterioration cost (DC) in the cycle 0; t1 ;

"
!#
Z l
Z t1

C4
C4 h
li

DC
at dt
al dt
I0 
I0  al t1 
T
T
2
0
l
"
#
b1
b2
b2

C 4 a alt1  c
al  c  c
l
;

 acb l t 1 

T
b 1b 2
b1
2

3:10

Shortage cost (SC) in the interval t1 ; T;

SC

C2
T

It dt

t1

C 2 al T  t1 2
:
:
T
2

3:11

Ordering cost(OC);

OC

C3
T

3:12

and the inventory holding cost (HC) in 0; t1 ;

C1
HC
T

Z l

It dt

t1

It dt :

3:13

To make the derivation of the cost function simple, we approximate the inventory depletion curve with a straight line. Similar treatment of the inventory depletion curve with linear approximation can be found in [9,11,12]. Thus, the inventory
holding cost per unit time is approximately;

HC

1 C1
: I0 t 1 :
2 T

3:14

See Appendix A for exact inventory holding cost from Eq. 3.13.
Hence, the total inventory cost per unit time, uT; t1 is given by:

uT; t1


C4 
l 1 C 1
C 2 alT  t 1 2 C 3
:
: I0 t 1
I0  al t 1 
2 T
T
2
2T
T

3:15

We assume t 1 KT; 0 < K < 1. This assumption seems reasonable since the length of the shortage interval is part of the cycle length. In addition, this restriction on values of t1 enhances the convexity of the inventory cost function in Eq. 3.15.
Substituting t 1 KT in (3.15), we get



C4 C1K
C 4 al 
l C 2 alT  KT2 C 3
I0 

KT 
2
T
T
2
2T
T
"
#



C 4 C 1 K alKT  cb1 al  cb2  cb2
l
b
 l KT  ac  1 ;


a
2
b 1b 2
T
b1
2

uT; K

3:16

C 4 al 
l C 2 al1  K2 T C 3
:
KT 

2
T
2
T

We now proceed to determine T and K optimally by treating them as decision variables. The Inventory cost per unit time,

uT; K being a function of two variables T and K has to be partially differentiated with respect to T and K separately and
then put equal to zero. This gives the necessary condition for minimizing the total inventory cost per time uT; K. That is;
"
#
@ u C 4 a alKTb 1KT  cb  KT  cb1 al  cb2  cb2 l2
b

 ac  1
2
b 1
b 1b 2
@T
2
T

C 1 alK 2
C 4 al2 C 2 al1  K2 C 3
aKT  cb  acb  1

 2 0;
2
2
2T 2
T

3:17

@u
C 4 alaKT  cb  acb  1
@K
"
#

C 1 a alKTb 1KT  cb KT  cb1 al  cb2  cb2
l
b

 l 2KT  ac  1

2
b 1
b 1b 2
2
 C 4 alT  C 2 al1  KT 0:

3:18


The solutions of Eq. 3.17 and Eq. 3.18, solving simultaneously, give the optimal values T and K which minimize the total
inventory cost per unit uT; K provided they satisfy the sufcient condition below.

9703

S.S. Sanni, W.I.E. Chukwu / Applied Mathematical Modelling 37 (2013) 96989706

"
l1 C 4 a

al

b 1T

 

4

h
b1
b
b
T 2 K  b 1K  T  bK  T   c
K  T   c   K  b 1K  T   c

 

 

2TK T b 1K T  c  K T  c

b1

#
i al  cb2  cb2 l2 acb  1


b 1b 2T 3
2T 3

C 1 aK 3 l
C al2 C
abK  T   cb1  4 3 33 ;
2
2T
T

l2

C 4 aal
T 2 b 1

T  b 1K  T  bK  T   c

b1

K  T   c  b 1K  T   c 

C 1 al
b1
abK 2 T  K  T   c
2

2K  K  T   c  2K  acb  1 C 2 al1  K  ;
l3 C 4 aK  labK  T   cb1
"
#
b1
b
b
C 1 a alK  b 1K  T  bK  T   c
K  T   c K  b 1K  T   c 
b


 2K lac  1  C 4 al
2
b1
 C 2 al1  K  ;
b1

l4 C 4 aT  labK  T   c
"
#
b1
b
b
C 1 a alT  b 1T  K  bK  T   c
K  T   c T  b 1K  T   c 
b


 2T lac  1 C 2 aT  l;
2
b1
0

where li s; i 1; 2 . . . 4, are: @ 2 u=@T 2 ; @ 2 u=@T@K; @ 2 u=@K@T and @ 2 u=@K 2 respectively.

The sufcient condition is:

l1 > 0;

l4 > 0;

l1 :l4  l2 :l3 > 0:

3:19

If l2 l3 , the condition reduces to:

l1 > 0;

l4 > 0;

l1 :l4  l2 > 0:

Remark 2. We can put (3.19) in equivalent form as; the total cost per unit time uT; K is minimized if its Hessian matrix
evaluated at T  ; K  is positive denite.
The total back-order quantity for the cycle is alT   t1 . Thus, the optimal order quantity , I is:

I I0 alT   t 1 :

3:20

Corollary 3.1. A change in the inventory ordering cost C 3 by the amount qT, where q is the ordering cost per unit of item produced, will leave the optimal ordering quantity unchanged.
Proof. Replacing C 3 in 3.16 with C 3 qT, we obtain;

#

"

C 4 C 1 K alKT  cb1 al  cb2  cb2
l
b
 l KT  ac  1
u T; K a


2
b 1b 2
T
b1
2
D

C 4 al 
l C 2 al1  K2 T C 3
q:
KT 

2
T
2
T

3:21

Differentiating 3.21 partially w.r.t T and K, we observed that @ uD =@T @ u=@T 0 and @ uD =@K @ u=@K 0. Hence, the
result follows and the proof is complete. h

4. Numerical example
Example 4.1. An example is chosen for numerical illustration which represents an inventory system with the following
data:
C 1 2.40 /unit/year, C 2 5=unit=year;C 3 100=order;C 4 8=unit;a 9000 units/year, l 2=3 year, a 0:002; b 20
and c 0:6.
The optimal solutions are found to be:
K  0:9907; T  1:893 years;t1 1:8754 years;I0 9346:8851 units;I 9452:5145 units and u 11566:60=year.

9704

S.S. Sanni, W.I.E. Chukwu / Applied Mathematical Modelling 37 (2013) 96989706

Table 1
Sensitivity analysis of the inventory model parameters.
Parameter

C1

C2

C3

C4

% Change

+50
+25
25
50
+50
+25
25
50
+50
+25
25
50
+50
+25
25
50
+50
+25
25
50
+50
+25
25
50
+50
+25
25
50
+50
+25
25
50
+50
+25
25
50
+20
25
50

% Change in
T

I0

u

1.61
0.81
+0.82
+1.65
0.00
0.00
0.00
+0.01
+1.05
+0.04
0.04
0.08
+1.05
+0.62
1.02
3.01
0.05
0.03
+0.05
+0.78
6.78
3.43
+3.86
+8.91
3.60
2.00
+2.64
+6.49
18.03
11.12
+20.15
+61.75
+6.55
+3.27
3.26
6.51
16.66
+33.33
+99.97

2.52
1.29
+1.34
+2.73
0.00
0.00
0.00
0.00
+0.10
+0.06
0.06
0.13
+1.71
+1.01
1.62
4.63
+49.87
+24.94
24.94
49.87
+27.58
+15.29
18.08
38.94
3.71
2.06
+2.71
+6.66
20.43
12.52
+22.14
+65.82
+6.25
+3.11
3.08
6.12
+0.003
0.008
+0.02

+27.08
+13.65
13.88
28.00
+.29
+0.14
0.15
0.29
+2.20
+1.44
1.44
2.89
+19.02
+9.56
9.78
20.07
+47.11
+23.57
23.56
45.97
+12.51
+7.59
10.21
23.20
0.91
0.51
+0.71
+1.82
16.29
9.29
+13.87
+35.51
0.77
0.52
+0.84
+2.05
+19.50
22.33
39.79

Example 4.2. Let: C 1 0:15 /unit/year, C 2 0.02 /unit/year, C 3 280=order, C 4 1.5/unit, a 15000 units/year,
year, a 0:001, b 8, c 0:4 and K 0:8.
The optimal solutions to the inventory problem are:
T  3:2416 years, t 1 2:5933 years, I0 18553:5546 units, I 23416:0230 units and u 1662.25/year.

l 1=2

We now proceed to test the responsiveness of the proposed model to changes in the model parameters using Example 4.2.
The sensitivity analysis is performed by changing the value of each parameter by 50%, 25%, 25%, 50%, taking one parameter at a time and keeping the remaining parameters unchanged. We assume that insensitive, moderately sensitive, and
highly sensitive imply percentage changes are 3 to 3, 20 to +20 and more respectively. This guide for categorizing sensitivity is used by [16].
It can be seen from Table 1 that the solutions T  ; I0 and u are highly insensitive to change in C 2 and C 3 while the solutions
are highly sensitive to changes in l and b. Furthermore, the initial optimal quantity I0 and the optimal inventory cost u
change by roughly the same percentage change in a.
5. Conclusion
We have presented an inventory model for three-parameter Weibull distribution deteriorating items with ramp-type demand rate. The proposed model is suitable for newly launched product with erratic demand pattern up to a point in time.
The Weibull distribution is often used for modeling duration data because it has exponential and Rayleigh distributions as
sub-models. In practice, the rate of deterioration of most items increases with time. Cooray [19] suggested that when modeling monotone hazard rates, the weibull distribution may be an initial choice because of its negatively and positively
skewed density shape. The location parameter, c, in the model depicts the shelf-life of the item under consideration; an
essential feature often ignored by most inventory modelers. Several intrinsic features of the model have demonstrated via

S.S. Sanni, W.I.E. Chukwu / Applied Mathematical Modelling 37 (2013) 96989706

9705

sensitivity analysis. In particular, the optimal quantity and the total inventory cost change with approximately the same percentage changes in a for the inventory problem in Example 4.2. Also, we give the necessary and sufcient conditions for the
optimal solutions to the model. The proposed model can be extended in a number of ways; it would be interesting to relax
the assumption of no discount quantity or to consider the parameters as fuzzy/stochastic fuzzy.
Addendum
Material below came to the attention of the authors at the time of revising this paper.
Note 1. Jalan et al [20] presented an EOQ model for items with two-parameter Weibull distribution deterioration and ramp
type demand. They allowed shortages in the inventory and obtained the economic order quantity via numerical technique.
Note 2. Giri et al [21] reconsidered the model in [20] by taking a ramp type demand pattern for fashionable products which
initially increases exponentially with time up to a point after which it becomes steady.
Note 3. Jain and Kumar[22] developed an EOQ model for inventory system that starts with shortage. They considered the
case where the time to deterioration of items follows three-parameter Weibull distribution, the demand rate is ramp type
function of time that increases exponentially up to a point then stays constant and they provided an analytical method for
reaching the optimal replenishment policy.
The present authors feel that inventory model that starts with shortage is appropriate for faulty inventory system which
experience unnecessary delay at the initial stage due to delay in supply, shortage of raw materials, shortage of labour, etc.
Such system rarely occurred in practice because of technological advancement and business competition. We are also of the
opinion that exponential time-varying demand rate is unrealistic because the demand of any product seldom experiences a
rate which is as high as exponential rate. Finally, Two-parameter Weibull distribution instantaneous rate function is appropriate for item with decreasing rate of deterioration only if the initial rate of deterioration is very high or it can be used for
system with increasing rate of deterioration only if the initial rate is negligible[see [10,11]].
Acknowledgements
We thank the referees for their valuable comments and suggestions which led to a substantial improvement of this paper.
The rst author would like to thank Yekini Shehu for helpful discussions.
Appendix A. Derivation of the total average holding cost
We derive the Total average holding cost per cycle directly from the inventory depletion curve as follows:

"Z
#
Z t1
l
C1
HC
It dt
It dt
T
0
l
" Z "
#
l
C1
1 2
att  cb1 at  cb2  cb2
b
b
I0 expfac g dt
a
t 1  at  c


2
b 1b 2
T
b1
0
"
#
#
Z t1
at1  cb1  t  cb1

alt 1  t1  a1  cb
dt
b1
l
" "
!#
C1
1 l3
l2 l  cb1
2
b1
b3
b3
lb 3l  c
a
a

 l  c
 c 

2 3
b 1b 2b 3
T
b1

a
b 1b 2b 3

lb 3l  cb2  l  cb3  cb3 


l  cb3  cb3  b 3lcb2  lI0 expfacb g
b 1b 2b 3
""
#
t1
l al  cb1 t1  l t1  cb2  l  cb2
al
 lt 1 

b1
b 1b 2
2
2
h
i
a
:
b 2t 1 t1  cb1  ll  cb1  t 1  cb2  l  cb2

b 1b 2


Appendix B. Derivation of the inventory level

Using the integrating factor(i.e. expfat  cb g) of the rst order linear differential equation of (3.1) to multiply both sides
of (3.1), we get

9706

S.S. Sanni, W.I.E. Chukwu / Applied Mathematical Modelling 37 (2013) 96989706

d
It expfat  cb g at expfat  cb g:
dt
Solving the above equation using the condition, I0 I0 , we obtain

Z t
Iteatcb  I0 eacb a
t1 at  cb  dt;
0
"
#
t2 att  cb1 at  cb2  cb2
atcb
I0 expfacb g;
a


Ite
b 1b 2
2
b1
"

"
#
#
t 2 att  cb1 at  cb2  cb2
b
) It a


I0 expfac g expfat  cb g;
b 1b 2
2
b1

0 6 t 6 l:

Note that we have used eatcb  1 at  cb on the integrand since 0 < a  1.

Similarly, the solutions to the differential Eqs. (3.2) and (3.3), after using the boundary conditions, give (3.5) and (3.6).
Hence, the inventory level at any time t 2 0; T is:

8h h
i
i
2
cb1
cb2 cb2
>
a t2 att
 atb1b2
I0 expfacb g expfat  cb g; 0 6 t 6 l;
>
b1
>
<h h
ii
at1 cb1 tcb1
It
l 6 t 6 t1 ;
expfat  cb g;
> al t 1  t
b1
>
>
:
t1 6 t 6 T;
alt  t 1 ;
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]

P.K. Gupta, D.S. Hira, Operations Research, S. Chand & Company LTD, New Delhi, 2002.
A.C. Hax, D. Candea, Production and Operations Management, Prentice-Hall, Englewood Cliffs, New Jersey, 1984. pp. 135.
T.M. Whitin, Theory of Inventory Management, Princeton University Press, New Jersey, 1957. pp. 6272.
H.M. Wagner, T.M. Whitin, Dynamic version of the economic lot size model, Manage. Sci. 5 (1958) 8996.
P.M. Ghare, G.F. Schrader, A model for exponentially decaying inventory, J. Ind. Eng. 14 (1963) 238243.
W.A. Donaldson, Inventory replenishmemt policy for a linear trend in demand: an analytical solution, Oper. Res. Q. 28 (1977) 663670.
E.A. Silver, A simple inventory replenishment decision rule for a linear trend in demand, J. Oper. Res. Q. 30 (1979) 7175.
M. Deb, K.A. Chuadhuri, A note on the heuristic for replenishment of trended inventories considering shortages, J. Oper. Res. Soc. 38 (1987) 459463.
R.P. Covert, G.C. Philip, An EOQ model for items with Weibull distribution deterioration, AIIE Trans. 5 (1973) 323326.
G.C. Philip, A generlized EOQ model for items with Weibull distribution deterioration, AIIE Trans. 6 (1974) 159162.
B.C. Chakrabarty, B.C. Giri, K.S. Chuadhuri, An EOQ model for items with Weibull distribution deterioration, shortages and trended demand: an
extension of Philips model, Comput. Oper. Res. 25 (1998) 649657.
S.K. Ghosh, K.S. Chaudhuri, An order-level Inventory model for a deteriorating item with Weibull distribution deterioration, time-quadratic demand
and shortages, Adv. Model. Optim. 6 (1) (2004) 2135.
S.S. Sanni, An economic order quantity inventory model with time dependent weibull deterioration and trended demand, M.Sc. thesis, University of
Nigeria, 2012.
B. Mandal, A.K. Pal, Ordel level inventory system with ramp type demand rate for deteriorating items, J. Interdisciplinary Math. 1 (1998) 4966.
K.S. Wu, L.Y. Ouyang, A replenishment policy for deteriorating items with ramp type demand rate, Proc. Natl. Sci. Counc. ROC (A) 24 (2000) 279286.
G.P. Samanta, J. Bhowmick, A deterministic inventory system with Weibull distribution deterioration and ramp-type demand rate, J. App. Stat. Anal. 3
(2) (2010) 92114 (Electronic).
F. Raafat, Survey of literature on continuous deteriorating inventory models, J. Opl. Res. Soc. 42 (1991) 2737.
R. Li, H. Lan, J.R. Mawhinney, A rewiew on deteriorating inventory study, J. Service Sci. Manage. 3 (2010) 117129.
K. Cooray, Generalization of the Weibull distribution: the odd Weibull family, Stat. Model. 6 (2006) 265277.
A.K. Jalan, B.C. Giri, K.S. Chaudhuri, An EOQ model for items with Weibull distribution deterioration, shortages and ramp type demand, Recent
Development in Operations Research, Narosa Publishing House, India, 2001. pp. 207223.
B.C. Giri, A.K. Jalan, K.S. Chaudhuri, Economic order quantity with Weibull distribution deterioration, shortage and ramp type demand, Int. J. Syst. Sci. 4
(2003) 237243.
S. Jain, M. Kumar, An EOQ model for items with ramp type demand, three parameter Weibull distribution deterioration and starting with shortage,
Yugoslav J. Oper. Res. 20 (2010) 249259.