# A Discrete Time Two Commodity Inventory System

DOI : 10.17577/IJERTV5IS010602

Text Only Version

#### A Discrete Time Two Commodity Inventory System

P. Senthil Kumar

School of Mathematics Madurai Kamaraj University Madurai – 625021, India

Abstract In this article, we consider a two-commodity inventory system under discrete time review. The demands for each commodity – i(i 1,2) arrive according to a independent

#### Bernoulli process. The maximum inventory level for the

i

i

ith (i = 1,2) commodity is fixed as S , i = 1,2 and the

reorder level as si , i = 1,2 . The ordering policy is defined as, when both the inventory levels are less than or equal to their respective reorder levels, we place an order for Qi (= Si si ), i = 1,2 units . The lead time distribution is assumed to be geometric. The demands that occur during the stock-out period are considered to be lost. Some system performance measures in the steady state are derived and the total expected cost rate under a suitable cost structure is calculated. The results are illustrated numerically.

Keywords D iscrete time; two -commodity; inventory system; joint order policy;

I. INTRODUCTION

One of the factors that contribute the complexity of the present day inventory system is the multitude of items stocked and this necessitated the multi-commodity systems. In dealing with such systems, in the earlier days models were proposed with independently established reorder points. But in situations were several product compete for limited storage space or share the same transport facility or are produced on (procured from) the same equipment (supplier) the above strategy overlooks the potential savings associated with joint ordering and, hence, will not be optimal. Thus, the coordinated, or what is known as joint replenishment, reduces the ordering and setup costs and allows the user to take advantage of quantity discounts.

Inventory system with multiple items have been subject matter for many investigators in the past. Such studies vary from simple extensions of EOQ analysis to sophisticated stochastic models. References may be found in [15, 2, 22, 24,

27, 17] and the references therein.

Kalpakam and Arivarignan [9] have introduced (s,

S) policy with a single reorder level s defined in terms of the total number of items in the stock. This policy avoids separate ordering for each commodity and hence a single processing of orders for both commodities has some advantages in situation where in procurement is made from the same supplies, items are produced on the same machine, or items have to be supplied by the same transport facility.

Krishnamoorthy and Varghese [10] have considered a two commodity inventory problem without lead time and with Markov shift in demand for the type of commodity namely commodity-1, commodity-2 or both commodity, using the direct Markov renewal theoretical results. Anbazhagan and Arivarignan [4, 5, 6, 7] have analyzed two commodity inventory system under various ordering policies. Yadavalli et al. [26] have analyzed a model with joint ordering policy and varying order quantities. Yadavalli et al. [27] have considered a two-commodity substitutable inventory system with Poisson demands and arbitrarily distributed lead time.

A two-commodity inventory system under continuous review is analyzed by Sivakumar [21]. They assumed that both the commodities are substitutable in the sense that at the time of zero stock, the other commodity is used to meet the demand. During the stock-out of both commodities an arriving demand entered the orbit of infinite size. They assumed constant retrial policy with exponential retrial time.

In all the above models, the authors assumed that the time axis is continuous. But, the discrete time systems, are more appropriate than their continuous time counterparts for modelling diverse productive processes, since the basic units in these systems are digital. In discrete time setting, it is assumed that the time axis is calibrated into epochs by small units and that all the events are deemed to occur only at these epochs. With the advent of fast computing devices and efficient transaction reporting facilities, such epochs with small gaps can be conveniently assumed so that events can occur at these epochs.

The analysis of discrete time queueing models has received considerable attention in the literature over the past years, in view of its applicability in the study of many computer and communication systems in which time is slotted ([23], [25]). An important application stems from the secondary and tertiary sector since, for example the current production systems of numerous factories operate on a discrete time basis where events can only happen at regularly spaced epochs.

In the case of inventory modelling under discrete times, the first paper was by Bar-Lev and Perry [8], who assumed that demands are non-negative integer valued random variables and items have constant life times. Lian and Liu [12] developed a discrete time inventory model with geometrically distributed inter-demand times, bulk demands

and constant life time for items. They assumed

(0, S)

• The reorder level for the commodity – i is fixed as

ordering policy, with instantaneous supply which clears any

si (1 si < Si ) with an ordering quantity for the

backlog and restores the stock to the maximum capacity S.

commodity – i is Q (= S s > s 1)(i = 1,2) items

This assumption helped them to have fixed life time for all items. They derived the limiting distribution of inventory level through matrix-analytic method.

Lian et al. [13] developed a discrete time inventory system with discrete PH-renewal process for (batch) demand time points and assumed discrete-PH-distribution for life time of items. They also assumed zero lead time and that unsatisfied demand were completely backlogged.

Abboud [1] studied a discrete inventory model for production inventory systems with machine breakdowns. They assumed that the demand and production rates were constant and that the failure and repair times of each item were independently distributed as geometric.

In this paper we have considered a discrete time two commodity inventory system with independent reorder levels where a joint order for both commodities is placed only when the levels of both commodities are less than or equal to their respective reorder levels. The rest of the paper is organized as follows: In section 2, the mathematical model is described. Section 3 is the central one. The steady state analysis of the model is presented in section 4. Somekey system

i i i i

when both inventory levels are less than or equal to their respective reorder levels. The requirement Qi > si 1(i = 1,2) ensures that after the replenishment the inventory levels of both commodities are above their respective reorder levels; otherwise it may not be possible to reorder (according to this policy) which leads to perpetual shortage. The lead time is assumed to be distributed as geometric distribution with parameter b(> 0) .

• The demands that occur during stock-out periods are considered to be lost.

We assume that all the above activities are occurred in the slot. Therefore more than one event may occur in the same slot. For mathematical clarity, we need to define the order of the events to be occurred, here, first the replenishment of order, then the demand for the commodity – 1 and finally the demand for the commodity – 2 will be satisfied.

3 Analysis

Let X denote the inventory level of commodity – 1 and Y

performance measures are derived and the total expected cost tt

rate is calculated in section 5. In section 6, some numerical results are presented to illustrate the effect of the parameters on several performance characteristics.

II . Model Description

We consider a two-commodity inventory system where the time axis is divided into intervals of equal length, called slots. The end points of the slots are called slot boundaries. The

denote the inventory level of commodity – 2 at time t . From the assumptions made on the input and output processes, it can be shown that the stochastic process

{(Xt ,Yt ) : t = 0,1,2,,} is a discrete time Markov chain with state space given by

E = {(i, k) : i = 0,1,2,, S , k = 0,1,2,, S }.

system is monitored at the slot boundaries. The maximum 1 2

storage capacity for the commodity – i is Si (i = 1,2) which is used to meet the demands. The demand is for single item per customer. We assume the following:

• The demand for the commodity – 1 arrives according to a

The transition probability function is defined as for

(i, k),( j,l) E ,

Bernoulli process with probability a1 . Thus a1

is the

p((i, k),( j,l)) = Pr[X

= j,Y = l | X = i,Y

= k].

probability that a demand occurs at a slot and

a (= 1 a ) ,

t 1

t 1 t t

1 1

is the probability that a demand does not occur in a slot.

The transition probability matrix P of this process,

• The demand for the commodity – 2 arrives according to a

Bernoulli process with probability a2 . Thus a2

is the

P = (

p((i, k),( j,l)) ),

(i, k),( j,l) E

probability that a demand occurs at a slot and

a2 (= 1 a2 ) ,

is the probability that a demand does not occur in a slot.

Hence, we have

p((i, k),( j,l)) =

a1a2 ,

j = i, l = k, (or)

i = s1 1, s1 2,, S1 k = 1,2,, S2

ba1 ,

j = i 1,

l = k,

i = 1,2,, s1 k = 0

j = i,

l = k,

i = 1,2,, s1

k = s2 1, s2 2,, S2

ba2 ,

j = i, l = k,

i = 0

k = 1,2,, s2

a a ,

1 2

j = i,

l = k 1, (or)

j = i,

i = s1 1, s1 2,, S1 k = 1,2,, S2

i = 1,2,, s1

ba ,

2

2

ba1a2 ,

j = i,

l = k 1,

j = i,

i = 0

k = 1,2,, s2

i = 1,2,, s1

l = k 1,

k = s2

1, s2

2,, S2

l = k,

k = 1,2,, s2

1

1

2

2

1

1

a a , j = i 1,

i = s 1, s 2,, S

b

b

1

1

2

2

a a ,

j = i,

i = 1,2,, s

l = k,

(or)

j = i 1,

k = 1,2,, S2

i = 1,2,, s

a a ,

1

1

2

2

1

1

l = k 1,

j = i 1,

k = 1,2,, s2

i = 1,2,, s

l = k,

1

k = s2 1, s2 2,, S2

b 1 2

l = k,

1

k = 1,2,, s2

a a , j = i 1,

i = s 1, s 2,, S

a a , j = i 1, i = 1,2,, s

1 2

1 1 1

b 1 2 1

a1 ,

l = k 1, (or)

j = i 1,

l = k 1,

j = i,

k = 1,2,, S2

i = 1,2,, s1

k = s2 1, s2 2,, S2

i = s1 1, s1 2,, S1

ba a ,

1 2

ba1a2 ,

l = k 1,

j = i Q1 , l = k Q2 ,

j = i Q1 ,

k = 1,2,, s2

i = 0,1,2,, s1

k = 0,1,2,, s2

i = 0,1,2,, s1

a ,

l = k,

j = i 1,

k = 0

i = s 1, s 2,, S

a ,

l = k Q2 1,

j = i Q 1,

k = 0,1,2,, s2

i = 0,1,2,, s

1

l = k,

1 1 1

k = 0

ba1 2

1 2

1 2

1

l = k Q2 ,

1

k = 0,1,2,, s2

a2 , j = i,

i = 0

ba a ,

j = i Q1 1,

i = 0,1,2,, s1

We define

< i >= ((i,0),(i,1),,(i, S2 )),i = 0,1,2,, S1. By

ordering the set of states as (< 0 >,< 1 >,, < S1 >), the

transition probability matrix P of the discrete time Markov chain can be conveniently expressed in a block partitioned form with entries,

A1,

A ,

A ,

2

j = i 1

j = i

i = s1 1, s1 2,, S1 i = s1 1, s1 2,, S1

a2 ,

l = k

k = s2 1, s2 2,, S2

A3 ,

A ,

j = i 1

j = i

i = 1,2,3,, s1

i = 1,2,3,, s

a ,

a ,

2

l = k 1

k = s2 1, s2 2,, S2

[P] = 4

1 a2b ,

l = k

k = 1,2,3,, s2

<i>< j >

A ,

j = i

i = 0

[ A5 ]kl =

5 a2b ,

l = k 1

k = 1,2,3,, s2

C1,

C2 ,

0,

j = i Q1 1 j = i Q1 otherwise.

i = 0,1,2,, s1

i = 0,1,2,, s1

b ,

0,

l = k otherwise.

k = 0

where

a1a2 ,

l = k

k = 1,2,3,, S2

[C ]

a1a2b,

a a b,

=

=

l = k Q2

l = k Q 1

k = 0,1,2,, s2

k = 0,1,2,, s

a a ,

l = k 1

k = 1,2,3,, S

1 kl 1 2 2 2

[ A ]

= 1 2

2 0,

otherwise.

1. kl

a ,

l = k

k = 0

1

0,

otherwise.

[C ]

a1a2b,

a a b,

=

=

l = k Q2

l = k Q 1

k = 0,1,2,, s2

k = 0,1,2,, s

a a ,

l = k

k = 1,2,3,, S

2 kl

1 2 2 2

1 2 2

0,

otherwise.

a a , l = k 1 k = 1,2,3,, S

[ A ]

= 1 2 2

2. kl

a ,

l = k

k = 0

It may be noted that the matrices

A , A , A , A , A ,C

and

1 1 2 3 4 5 1

0,

a1a2 ,

otherwise.

l = k

k = s2 1, s2 2,, S2

C2 are square matrices of size S2 1.

4 Steady State Analysis

It can be seen from the structure of the transition probability matrix P , that the discrete time Markov chain

a a ,

l = k 1

k = s 1, s 2,, S

{(Xt ,Yt ),t = 0,1,2,,} on the finite state space (E) is

1 2

2 2 2

irreducible. Hence, the limiting distribution, exits and it is

[ A3 ]kl

a a b ,

= 1 2

a a b ,

l = k

l = k 1

k = 1,2,3,, s2

k = 1,2,3,, s

2

2

defined as

1 2

= lim Pr[ X = i,Y = k | X , L ]

a1b ,

l = k

k = 0

(i,k ) t t 0 0

t

0,

a1a2 ,

otherwise.

l = k

k = s2 1, s2 2,, S2

exists and is independent of the initial state. We group the probabilities (i,k ) as follows:

1

1

(i) = (i,0),(i,1),,(i,S ) ,i = 0,1,, S1

a a ,

l = k 1

k = s 1, s 2,, S

2

1 2

2 2 2

and = (0) ,(1) ,,(S ) .

[ A4 ]kl

a a b ,

= 1 2

a a b ,

l = k

l = k 1

k = 1,2,3,, s2

k = 1,2,3,, s

Then, the limiting probability distribution satisfies the

1 2

2 following equations

a1b ,

0,

l = k

otherwise.

k = 0

P =

and e = 1.

1

1

The first equation of the above yields the following set of equations:

and(Q 1) can be obtained by solving

s1

2 1

2 1

s 1

(s j )

A

A = ,

i = 0

[(I A ) A

1] 1 D C A 1[(I (3A)

) A 1] 1

(i1) 3

1. 5

(i)

(Q1 1)

j =0

j 1 1 2 1

(i1) A3 (i) A4 = (i) ,

i = 1,2,, s1 1

s1 1

( s 1k )

(4)

• D C A 1[(I A ) A

1] 1

(I A )

A A = ,

i = s

k 2 1 2 1

(25) = 0

(i1) 1 (i) 4

(i) 1

k =0

1 s1 1

1

A

A = ,

i = s

1, s

2,,Q 2

[ A1(I A2 ) ] [ A1(I A4 )

]C2

(6

(i1) 1

(i) 2

(i) 1 1 1

)

and

1

1

(i1) A1 (i) A2 (iQ 1)C1 = (i) ,

i = Q1 1

s1

Q1 2

(7)

1

1

Q 1i

A

A

C

C = ,

D0 Di

[ A1 (I A2 ) ] 1 I

(i 1) 1

1. 2

(i Q1 1) 1

(i Q1 ) 2

(i)

(Q 1)

i =1

i = s1 1

i = Q , Q 1,, S 1

1

(8)

1

1

1

1 1 1

[(I A2 ) A1

] D0C1 A1

1

1

(i) A2 (iQ )C2 = (i) ,

i = S1

[(I A

) A 1 ]i(Q11)

(9)

The equations (expect (7)) can be recursively solved to get

2 1

S1

S1

j =0

j =0

D C

D C

A 1[(I A

A 1[(I A

) A 1 ]i(Q1 j )

) A 1 ]i(Q1 j )

iQ1

=

=

,

,

i = 0,1,, S .

i = 0,1,, S .

(i) (Q11) i 1

j 1 1

2 1(10)

where

Di ,

i = 0,1,, s1,

i=Q11

A (I A )1 Q1 1i ,

iQ11

1 2

• D C

A 1[(I A

) A 1 ]i(Q11 j ) e = 1

i = s

1, s

2,,Q 2

k 2 1 2 1

I ,

1

i = Q1

1 1

1,

k =0

5 System Performance Measures

= 2 1 0 1 1

= 2 1 0 1 1

[(I A ) A 1 ] D C A 1,

i

[(I A ) A 1 ]i (Q1 1)

i = Q1

In this section, we derive some importance system performance measures.

1. Expected Inventory level

2 1

i Q1

i (Q j )

Let

denote the expected inventory level for the

D C A

1[(I A ) A 1 ] 1 I1

j = 0

j 1 1 2 1

commodity – 1 in the steady state.

i Q1 1

S1 S2

D C

A 1[(I A ) A 1 ]i (Q1 1 j )

I = i(i,k ).

k = 0

k 2 1 2 1 ,

1

i=1 k =0

i = Q

1,Q 2,, S .

Let

2

I

I

denote the expected inventory level for the

1 1 1

with

I

I

commodity – 2 in the steady state.

D = A (I A )1 s1 1A (I A )1 A (I A )1 s1 1

S1 S2

0 1 2

1 4 3 4

I = k(i,k ).

A (I A )1 and

2

i=0 k =1

3 5

D = A (I A )1 s1 1A (I A )1 A (I A )1 s1 i ,

2. Expected Reorder Rate

i 1 2 1 4 3 4

i = 1,2,, s1

Let R denote the expected reorder rate in the steady state.

s1 s2

function TC analytically. Hence a detailed computational study of the expected cost rate function is carried out in the

R = a2(i,s 1) a1( s 1,k ) a1a2( s 1,s 1).

next section.

(13)

i=0

2 1 1 2

k =0

1. 6 Numerical Analysis

3. Expected Shortage Rate

1

1

Let SR denote the expected shortage rate of commodity – 1

To study the behaviour of the model developed in this work, several examples were performed and the set of representative results are shown here. Although we have not

in the steady state.

shown the convexity of TC(s1, s2 ), our experience with

S2 considerable numerical examples indicate that the function

1

1

SR = a1(0,k ).

TC(s , s ), is convex.

(14)

Let SR

k =0

denote the expected shortage rate of commodity – 2

1 2

A three dimensional plot of

TC(s1, s2 ) is

2 presented in Figure 1. We use simple numerical search

in the steady state.

S1

procedure to get the optimal values of TC, s1

and

s2 (say

SR 2 (i,0) 1 2

= a .

2

TC*, s* and s* respectively). The minimu(1m5) value of

i=0

TC = 0.829224 is obtained at (s*, s* ) = (12,4) .

1 2

2

2

4. Total Expected Cost Rate

The long-run total expected cost per unit time for this system in the steady state is given by

TC(S1, S2 , s1, s2 ) = ch I

• ch I

• csR cr SR

• cr SR

where

1 1 2 2

1 1 2

 ch 1 The inventory carrying cost of commodity – 1 per unit item per unit time. ch 2 The inventory carrying cost of commodity – 2 per unit item per unit time. cs Setup cost per order. cr 1 Shortage cost of commodity – 1 per unit item per unit time. cr 2 Shortage cost of commodity – 2 per unit item per unit time.

By putting the values of s ' from the above measures of system performance, we obtain TC(S1, S2 , s1, s2 ) as

Figure 1: A three dimensional plot of total cost rate per unit time.

We have studied the effect of varying the costs and other system parameters on the optimal values and some of our

S1 S2 S1 S2

results are presented in Tables 1 to 12. The lower entry in

TC(S1, S2 , s1, s2 ) = ch i(i,k ) ch

k(i,k ) each cell corresponds to the total optimal cost value TC*

s1

• c a

1

i=1 k =0

s2

• a

• a a

2

i=0 k =1

1

1

and the upper entries correspond to the local optima

S * respectively.

2

2

S * and

s 2

(i,s2

1)

1 ( s1 1,k )

1 2 ( s1 1,s2

1)

Example 1: We start by examining the effect of the system

i =0

k =0

parameters namely, demand for commodity-1 with the probability, a1 ,demand for commodity- 2 with the

S2

• c a

c

S1

a

probability, a2 and lead time distributions success

r1

1 (0,k )

r2

2 (i,0)

probability, b , on the optimal values

(S*, S* )

and the

k =0

i=0 1 2

Due to the complex form of the limiting distribution, it is difficult to discuss the qualitative behaviour of the cost

corresponding optimal cost observe the following:

TC* . From tables 1 to 3, we

• The total expected cost rate increases when REFERENCES

each of a1 and a2 increase and it decreases with the

1. Abboud, N. E., (2001), A Discrete-Time Markov Production-

increase in b .

1

1

• As is to be expected, S * increase and

2

2

S * decrease with

Inventory Model With Machine Breakdowns, Computers & Industrial Engineering, 39, 95 – 107.

2. Agarwal,V., (1984), Coordinated order cycles under joint

* *

1

1

2

2

the increase in a1 . Similarly, S decrease and S increase

replenishment multi-item inventories, Naval Logistic Research Quarterly, 131136.

with the increase in a2 . This is because, if

a1 and a2

3. Alfa, A. S., (2002), Discrete Time Queues and Matrix-Analytic

increases, then more demands occur. To avoid frequent stock out we have to maintain larger inventory.

1

1

2

2

• S * and S * decrease with the increase in b . Because if the lead time is small, a smaller stock would be preferable.

Example 2: Next, we study the impact of costs, namely, the

Methods, Top, 10(2), 147 – 185.

4. Anbazhagan, N. and Arivarignan, G., (2000), Two-Commodity continuous review inventory system with coordinated reorder policy, International Journal of Information and Management Sciences, 11(3), 19- 30.

5. Anbazhagan, N. and Arivarignan, G., (2001), A two- commodity coordinated inventory system with renewal

holding cost of first commodity

c , the holding cost of

h

h

1

demands, Electronnoe modelirovanie, 23(2), 516 (in Russian).

6. Anbazhagan, N. and Arivarignan, G., (2001), Analysis of Two

second commodity

c , the setup cost

h

h

2

cs , the shortage cost

-Commodity Markovian Inventory system with lead time, The Korean Journal of Computational and Applied Mathematics, 8(2), 427 – 438.

r

r

of first commodity c

1

and the shortage cost of second

7. Anbazhagan, N. and Arivarignan, G., (2003), Two-Commodity Inventory system with Individual and Joint Ordering Policies,

r 1 2

r 1 2

commodity c on the optimal values (S*, S* ) and the

2

corresponding total expected cost rate TC*. From tables 4 to 12, we observe the following:

1 2 1 2

1 2 1 2

• The total expected cost rate increases with the increase in each of the costs ch , ch , cs , cr and cr .

International Journal of Management and Systems, 19(2), 129 – 144.

8. Bar-Lev, S. K. and Perry, D. (1989), A Discrete Time Markovian Inventory Model for Perishable Commodities, Stochastic Analysis and Applications, 7, 243 – 259.

9. Kalpakam, S. and Arivarignan, G., (1993), A coordinated multi-commodity (s, S) inventory system, Mathematical and Computer Modelling, textbf18, 6973.

10. Krishnamoorthy, A. and Varghese, T. V., (1994), A two

h

h

• As each of holding costs c

1

and c

h

h

2

increases,

commodity inventory problem, International Journal of Information and Management Sciences, 3, 55 – 70.

the optimal values S * and S * are decreasing. This is

11. Latouche, G. and Ramaswami, V., (1999), Introduction to

1 2 matrix analytic methods in stochastic modeling,. ASA SIAM,

because if the holding cost is high, more cost is required to maintain the large inventory. To avoid this, we would prefer to maintain small amount of stock.

• As the setup cost cs increases, the optimal values

1 2

1 2

S * and S * are increasing. If the setup cost is high, we should hold more items in the stock.

12. Lian, Z. and Liu, L., (1999), A Discrete-Time Model for Perishable Inventory Systems, Annals of Operations Research, 87, 103 – 116.

13. Lian, Z., Liu, L. and Neuts, M. F., (2005), A Discrete-Time Model For Common Lifetime Inventory Systems, Mathematics of Operations Research, 30(3), 718 – 732.

• As the shortage cost

increases, the optimal value *

r

r

S

S

c

c

1 1

1. Liu, L. and Yang, T., (1999), An (s,S) random lifetime

inventory model with a positive lead time, European Journal of

S

S

S

S

S

S

2

2

increases and the * decreases. Similarly the shortage cost

Operations Research, 113, 52 – 63.

2

2

2. Miller, B. L., (1971), A multi-item inventory model with joint

r

r

1

1

c increases, the optimal value

2

increases.

* decreases and the *

probability back-order criterion, Operations Research, 1467 1476.

Table 1: Effect of a1 and a2 on the optimal values

 a 1 a2 0.6 0.62 0.64 0.66 0.68 0.7 0.5 36 32 36 31 37 31 38 31 39 31 40 31 0.596968 0.602735 0.608424 0.614051 0.619615 0.625116 0.52 35 32 36 32 37 32 38 32 38 31 39 31 0.601644 0.607352 0.613006 0.618606 0.624057 0.629384 0.54 35 33 36 33 36 32 37 32 38 32 39 32 0.606411 0.612075 0.617655 0.623039 0.628377 0.633667 0.56 35 34 35 33 36 33 37 33 38 33 38 32 0.611276 0.616877 0.62224 0.627569 0.63286 0.638019 0.58 35 35 35 34 36 34 37 34 37 33 38 33 0.616228 0.621672 0.626976 0.632251 0.63737 0.642355 0.6 35 35 35 34 36 34 37 34 37 34 38 34 0.620631 0.62609 0.631466 0.636824 0.641995 0.64691

Table 2: Effect of a1 and b on the optimal values

 a 1 b 0.64 0.66 0.68 0.7 0.72 0.74 0.2 37 35 37 34 38 34 39 34 40 34 41 34 0.637157 0.642385 0.647428 0.652465 0.6575 0.662533 0.3 36 34 37 34 37 34 38 34 38 33 39 33 0.631466 0.636824 0.641995 0.64691 0.651736 0.656385 0.4 35 34 36 34 37 34 38 34 38 33 39 33 0.627004 0.632242 0.637487 0.642732 0.647956 0.653081 0.5 35 34 36 34 37 34 37 34 37 33 38 33 0.622057 0.62754 0.633048 0.638513 0.643441 0.648374 0.6 35 34 35 34 36 34 36 33 37 33 38 33 0.617222 0.622643 0.627774 0.632791 0.63789 0.643072 0.7 34 34 34 33 35 33 36 33 37 33 38 33 0.611996 0.617102 0.622171 0.627354 0.632645 0.63804

Table 3: Effect of a2 and b on the optimal values

 a2 b 0.55 0.56 0.57 0.58 0.59 0.6 0.1 39 36 38 36 38 36 38 37 38 37 38 38 0.649383 0.652975 0.65673 0.660404 0.664113 0.667944 0.2 36 34 36 34 36 35 36 35 35 35 35 35 0.615055 0.617307 0.619912 0.621936 0.624798 0.626548 0.3 35 33 35 34 35 34 35 35 35 35 35 35 0.608965 0.611276 0.613486 0.616228 0.618129 0.620631 0.4 34 33 34 33 34 34 34 34 34 34 34 35 0.603423 0.605971 0.60882 0.610994 0.613707 0.616184 0.5 34 33 34 33 34 33 34 34 33 34 33 34 0.597483 0.600673 0.603083 0.605841 0.608789 0.611194 0.6 33 32 33 33 33 33 33 33 33 34 33 34 0.591711 0.594621 0.597293 0.600607 0.603158 0.60599

h

h

Table 4: Effect of c

1

and c

h

h

2

on the optimal values

 c p c p 0.01 0.011 0.012 0.013 0.014 0.015 0.01 39 36 38 35 38 35 36 34 36 34 36 34 0.658962 0.671693 0.684347 0.695961 0.706759 0.717556 0.011 38 35 38 35 37 34 36 34 35 33 35 33 0.681807 0.694461 0.706773 0.718256 0.728977 0.739367 0.012 37 34 37 34 37 34 35 33 35 33 35 33 0.704464 0.716751 0.729038 0.74038 0.750771 0.761162 0.013 37 34 36 33 36 33 35 33 35 33 34 32 0.726728 0.738965 0.750891 0.762175 0.772565 0.782552 0.014 36 33 36 33 36 33 34 32 34 32 34 32 0.7488 0.760726 0.772652 0.783876 0.793861 0.803847 0.015 36 33 35 32 35 32 34 32 34 32 33 31 0.770561 0.782485 0.794057 0.80517 0.815156 0.824974

h

h

Table 5: Effect of c

1

and cs on the optimal values

 c p c s 0.01 0.011 0.012 0.013 0.014 0.015 6 34 30 33 30 32 29 32 29 32 29 31 28 0.559854 0.572914 0.583524 0.594086 0.604649 0.615116 7 36 32 34 31 34 31 33 30 33 30 32 30 0.5878 0.600346 0.611571 0.622726 0.633614 0.644131 8 37 33 36 33 35 32 35 32 34 32 33 31 0.613436 0.625831 0.637443 0.649015 0.660361 0.670101 9 37 34 37 34 36 33 36 33 35 33 34 32 0.63714 0.649427 0.661682 0.673608 0.684192 0.694508 10 39 36 38 35 38 35 36 34 36 34 36 34 0.658962 0.671693 0.684347 0.695961 0.706759 0.717556 11 40 37 39 36 38 36 37 35 37 35 37 35 0.679541 0.692795 0.70546 0.717021 0.728226 0.739431

h

h

Table 6: Effect of c

1

and c

r

r

1

on the optimal values

 c p c r1 0.011 0.012 0.013 0.014 0.015 0.016 0.7 37 36 36 36 36 36 35 36 34 35 34 35 0.652792 0.663201 0.672823 0.682189 0.690589 0.698903 0.8 38 36 38 36 37 36 36 35 35 35 35 35 0.665074 0.676875 0.687695 0.697956 0.707688 0.716889 0.9 39 36 38 35 37 35 37 35 36 35 35 34 0.674766 0.687933 0.699703 0.711105 0.721956 0.732028 1.0 39 35 38 35 38 35 37 35 36 34 36 34 0.682774 0.696613 0.709516 0.722362 0.733689 0.744693 1.1 39 35 39 35 38 35 37 34 37 34 36 34 0.68885 0.703775 0.718196 0.730973 0.743522 0.755986 1.2 40 35 39 35 38 34 38 34 37 34 37 34 0.694616 0.70985 0.724622 0.739288 0.752122 0.764671

h

h

Table 7: Effect of c

1

and c

r

r

2

on the optimal values

 c p c r2 0.006 0.007 0.008 0.009 0.01 0.011 3 35 37 34 37 33 36 32 35 32 35 31 34 0.562557 0.577559 0.591813 0.605683 0.619037 0.632065 3.5 35 38 34 37 33 36 32 35 32 35 31 35 0.564629 0.579275 0.59347 0.607279 0.620633 0.632951 4 35 38 34 37 33 36 32 35 32 36 31 35 0.566403 0.58099 0.595126 0.608875 0.621751 0.633654 4.5 35 38 34 37 33 36 32 36 32 36 31 35 0.568176 0.582706 0.596783 0.610192 0.622483 0.634356 5 35 38 34 37 33 37 32 36 32 36 31 35 0.569949 0.584421 0.598219 0.610924 0.623215 0.635059 5.5 35 38 34 38 33 37 32 36 32 36 31 35 0.571723 0.585905 0.598981 0.611656 0.623946 0.635761

h

h

Table 8: Effect of c

2

and cs on the optimal values

 c p c s 0.01 0.011 0.012 0.013 0.014 0.015 6 34 30 33 29 33 29 32 28 32 28 31 27 0.559854 0.579678 0.598989 0.618152 0.636849 0.655391 7 36 32 35 31 34 30 34 30 33 29 33 29 0.5878 0.60832 0.6287 0.648599 0.668356 0.687667 8 37 33 36 32 36 32 35 31 35 31 34 30 0.613436 0.634927 0.655946 0.676762 0.697229 0.717445 9 37 34 37 34 36 33 37 33 36 32 36 32 0.63714 0.659405 0.681352 0.702948 0.724093 0.745113 10 39 36 38 35 37 34 37 34 36 33 36 33 0.658962 0.681807 0.704464 0.726728 0.7488 0.770561 11 40 37 39 36 39 36 38 35 37 34 37 34 0.679541 0.703039 0.72631 0.749101 0.771788 0.794052

h

h

Table 9: Effect of c

2

and c

r

r

1

on the optimal values

 c p c r1 0.012 0.013 0.014 0.015 0.016 0.017 0.75 35 33 34 32 33 31 33 31 33 31 32 30 0.655662 0.675952 0.696178 0.715965 0.735751 0.755042 0.8 35 33 34 32 33 31 33 31 33 31 32 30 0.661314 0.681624 0.701871 0.721658 0.741445 0.760759 0.85 35 32 35 32 34 31 34 31 33 30 33 30 0.666932 0.68717 0.707301 0.727034 0.746706 0.765932 0.9 35 32 35 32 34 31 34 31 33 30 33 30 0.671257 0.691494 0.711583 0.731316 0.750941 0.770167 0.95 35 32 35 32 34 31 34 31 34 30 34 30 0.675581 0.695819 0.715865 0.735598 0.755173 0.774125 1 36 32 36 32 35 31 35 31 34 30 34 30 0.679587 0.699618 0.719319 0.738814 0.757917 0.776868

h

h

Table 10: Effect of c

2

and c

r

r

2

on the optimal values

 c p c r2 0.012 0.013 0.014 0.015 0.016 0.017 2.5 36 32 36 32 35 31 35 31 34 30 34 30 0.679931 0.699962 0.71973 0.739225 0.758413 0.777365 3 35 32 35 32 35 31 35 31 34 30 34 30 0.680019 0.700256 0.720141 0.739636 0.75891 0.777861 3.5 35 32 35 32 34 31 35 31 34 30 34 30 0.680075 0.700313 0.720338 0.740046 0.759406 0.778358 4 35 32 35 32 34 31 34 31 33 30 34 30 0.680132 0.70037 0.720402 0.740135 0.7597 0.778854 4.5 35 32 35 32 34 31 34 31 33 30 33 30 0.680189 0.700427 0.720466 0.740199 0.759772 0.778999 5 35 32 35 32 34 31 34 31 33 30 33 30 0.680246 0.700484 0.72053 0.740263 0.759845 0.779071

1

1

Table 11: Effect of cs and cr

on the optimal values

 cs c r1 6 7 8 9 10 11 0.5 32 31 33 33 34 35 35 36 35 37 36 38 0.520918 0.545697 0.568678 0.589853 0.609871 0.628745 0.55 32 31 34 33 35 35 36 36 36 37 37 38 0.528152 0.553657 0.577177 0.598884 0.61914 0.638388 0.6 33 31 35 33 35 34 36 36 37 37 38 38 0.534011 0.560708 0.584562 0.60691 0.627499 0.6472 0.65 34 31 35 33 36 34 37 36 38 37 38 38 0.539671 0.566374 0.591075 0.613847 0.635036 0.655015 0.7 34 31 35 32 36 34 37 35 38 37 39 38 0.543823 0.571555 0.596722 0.620002 0.641799 0.661972 0.75 34 31 35 32 37 34 38 35 39 37 40 38 0.547975 0.575761 0.601546 0.625613 0.647485 0.668296

2

2

Table 12: Effect of cs and cr

on the optimal values

 cs c r2 9 10 11 12 13 14 2.5 34 29 35 30 36 31 37 32 38 33 38 34 0.587972 0.608885 0.628377 0.646635 0.663832 0.679977 3 34 29 35 30 36 31 37 32 38 33 38 34 0.588523 0.609261 0.628637 0.646818 0.663964 0.680007 3.5 34 29 35 30 36 31 37 32 37 33 38 34 0.589075 0.609637 0.628898 0.647002 0.664033 0.680036 4 34 29 35 30 36 31 37 32 37 33 38 34 0.589626 0.610012 0.629158 0.647186 0.664072 0.680065 4.5 34 29 35 30 36 31 36 32 37 33 38 34 0.590178 0.610388 0.629418 0.647304 0.664111 0.680095 5 33 29 34 30 35 31 36 32 37 33 38 34 0.590556 0.610735 0.629606 0.647357 0.66415 0.680124
3. Neuts, M.F., (1981), Matrix-geometric solutions in stochastic models: An algorithmic approach, Johns Hopkins University Press, Baltimore.

4. Ohno, K. and Ishigaki, T., (2001), A multi-item continuous review inventory system with compound Poisson demands, Mathematical Methods of Operations Research, 53, 147165.

5. Silver, E. A., (1974), A control system for coordinated inventory replenishment, International Journal of Production Research, 12, 647671.

6. Sivakumar, B., Anbazhagan, N. and Arivarignan, G., (2005), A two-commodity perishable inventory system, ORiON, 21(2), 157 – 172.

7. Sivakumar, B., Anbazhagan, N. and Arivarignan, G., (2005), Two-commodity continuous review perishable inventory system, International Journal of Information and Management Sciences, 17(3), 47 – 64.

8. Sivakumar, B., (2008), Two-commodity inventory system with retrial demands, European Journal of Operational Research, 187, 70 – 83.

9. Srinivasan, S. K. and Ravichandran, N., (1994), Multi-item

(Eds.), (Reviewed edited), Proceedings of the Conference of Stochastic Models, Optimization and Computer Application, Wiley Eastern Limited, 226236.

10. Takagi, H. (1993), Queueing analysis: A foundation of performance evaluation, Discrete-Time Systems, J.Math.Phy.Sci., 3, North-Holland, Amsterdam.

11. Thomstone, T. M. and Silver, E. A., (1975), A coordinated inventory control system, International Journal of Production Research, 581602.

12. Woodward, M. E. (1994), Communication and computer networks: Modelling with Discrete-time queues, IEEE Computer Society Press, Los Alamitos, CA.

13. Yadavalli, V. S. S., Anbazhagan, N. and Arivarignan, G., (2004), A two-commodity continuous review inventory system with lost sales, Stochastic Analysis and Applications, 22, 479 497.

14. Yadavalli, V. S. S., Van Schoor, C. De W. S. and Udayabaskaran., (2006), A substitutable two-product inventory system with joint-ordering policy and common demand, Applied Mathematics and Computation, 172 (2), 12571271.

(S, s)

inventory model with Poisson demand, general lead

time and adjustable reorder size, in: G.V. Krishna Reddy et al.