Int. J.

Production Economics ] (]]]]) ]]] ]]]

Contents lists available at ScienceDirect

Int. J. Production Economics

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

Optimal MRP parameters for a single item inventory with random replenishment lead time, POQ policy and service level constraint
Mohamed-Aly Louly a, Alexandre Dolgui b,n
FARCAMT, King Saud University, College of Engineering, Industrial Engineering Department, P.O. Box 800, Riyadh 11421, Saudi Arabia Ecole des Mines de Saint-Etienne, Centre for Industrial Engineering and Computer Science, Laboratory for Information Science and Technology, 158 cours Fauriel, 42023 Saint-Etienne Cedex 2, France
b a

a r t i c l e i n f o
Article history: Received 15 December 2009 Accepted 7 February 2011 Keywords: Supply planning Stochastic lead time Periodic Order Quantity MRP parameterization

abstract
This study deals with Material Requirement Planning (MRP) software parameterization under uncertainties. The actual lead time has random deviations; so it can be considered as a random variable. MRP approach with Periodic Order Quantity (POQ) policy is considered. The aim is to nd the optimal MRP time phasing corresponding to each periodicity of the POQ policy. This is a crucial issue in supply planning with MRP approach because inappropriate planned lead times under lead time uncertainties invariably lead to large and costly inventories or insufcient customer service levels. The proposed model and algorithms minimize the sum of the setup and holding costs while satisfying a constraint on the service level. Our approach does not need to employ the commonly used normal probability distributions. Instead, its originality is in nding a closed form of the objective function, valid for any probability distribution of the actual lead times. & 2011 Elsevier B.V. All rights reserved.

1. Introduction Effective replenishment is a crucial problem in supply planning. An inadequate inventory control policy leads to overstocking or stockout situations. In the former, the generated inventories are expensive and in the latter there are shortages and penalties due to unsatised customer demands. Material Requirements Planning (MRP) is a commonly accepted approach for replenishment planning in major companies (Axs ater, 2006). The MRP-based software tools are accepted readily. Most industrial decision makers are familiar with their use. The practical aspect of MRP lies in the fact that this is based on comprehensible rules, and provides cognitive support, as well as a powerful information system for decision making. Some instructive presentations of this approach can be found in Baker (1993), Sipper and Buln (1998), Zipkin (2000), Axs ater (2006), Tempelmeier (2006), Dolgui and Proth (2010) and Graves (2011). Nevertheless, MRP is based on the supposition that both demand and lead time are deterministic. However, most production systems are stochastic. For example, a random lead time can be explained by the variability of actual supplier load (when a supplier furnishes several clients, its load depends on the timing

Corresponding author. Tel.: +33 4 77 42 01 66; fax: +33 4 77 42 66 66. E-mail addresses: louly@ksu.edu.sa (M.-A. Louly), dolgui@emse.fr (A. Dolgui). URL: http://www.emse.fr/~dolgui (A. Dolgui).

of all client orders, if total demand outstrips production capacity, the lead time increases). There are many other external factors increasing randomness of lead times: outsourced production overseas can introduce some randomness via shipping perturbations, the orders might not arrive by the due date because of work stoppage or delays attributable to the weather (Graves, 2011). Additional random factors and unpredictable events such as machine breakdowns, absenteeism, other random variations of capacity can cause deviations in actual lead times from planning ones (Koh and Saad, 2003; Chaharsooghi and Heydari, 2010). Therefore, as aforementioned the deterministic assumptions of MRP can be often too restrictive. As shown in Whybark and Wiliams (1976), Ho and Lau (1994), Molinder (1997) and Chaharsooghi and Heydari (2010), lead time is a principal factor foreseeing production and lead time randomness affects seriously ordering policies, inventory levels and customer service levels. Thankfully, the MRP approach can be tailored to uncertainties by searching optimal values for its parameters (Buzacott and Shanthikumar, 1994; Hegedus and Hopp, 2001; Koh and Saad, 2003; Inderfurth, 2009; Mula and Poler, 2010). An adequate choice of these parameters increases the effectiveness of MRP techniques. Thus, one of the essential issues for companies in industrial situations is MRP parameterization. This is commonly called MRP offsetting under uncertainties. There are several MRP parameters: planned lead time, safety stock, lot-sizing rule, freezing horizon, planning horizon, etc. There exist extensive publications concerning safety stock

0925-5273/$ - see front matter & 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.ijpe.2011.02.009

Please cite this article as: Louly, M.-A., Dolgui, A., Optimal MRP parameters for a single item inventory with random replenishment lead time, POQ policy and service level.... International Journal of Production Economics (2011), doi:10.1016/j.ijpe.2011.02.009

M.-A. Louly, A. Dolgui / Int. J. Production Economics ] (]]]]) ]]] ]]]

calculation for random demand of nished products (Porteus, 1990; Lee and Nahmias, 1993). In contrast, certain parameters seem not to be sufciently examined as, for example, planned lead time (differences between due dates and release date). Optimal parameterization of most used lot-sizing rules is also an open issue. If actual lead time is random, the planned lead time can contain safety lead time, i.e. the planned lead time is calculated as the sum of the forecasted (or contracted) and safety lead times. The latter should be formulated as a trade-off between overstocking and stockout while minimizing the total cost. The search for optimal value of safety lead time, and, consequently, for planned lead time, is a crucial issue in supply planning with the MRP approach. The problem of planned lead times optimization, when safety lead times are used, has been given scant attention in the literature. In practice, often average values or percentiles of probability distributions of actual lead times are used. Nevertheless, a longer than necessary planned lead time creates excessive work in progress. Perhaps of special interest, Graves (2011) in his chapter Uncertainties and Production Planning of the Handbook of Production Planning and Inventories in the Extended Enterprise considers that there is a great opportunity for developing decision support to help planners in understanding the trade-offs and in setting these parameters in a more scientic way. This is one of motivations for this paper where we propose a decision support model for optimal MRP time phasing for each periodicity of the POQ policy. The proposed model and algorithms minimize the sum of the setup and holding costs while satisfying a constraint on the service level. Our approach does not need to employ the commonly used normal probability distributions. Instead, its originality is in nding a closed form of the objective function, valid for any probability distribution of the actual lead times.

Periods (time buckets) 1 40 Gross requirements Projected On-hand 100 60 Planned Order receipts Planned Order Releases

2 35 25 85

3 65 45 85

4 45 0 75

5 20 55 75

6 55 0 80

7 45 35 80

8 35 0

Fig. 1. An example of MRP table for POQ policy.

not possible (transportation constraints,y) or too costly (because of an expensive setup,y), so another lot-sizing rule should be used. Here, we consider the Periodic Order Quantity (POQ) policy that consists of grouping orders for p consecutive periods. In this case, the planned release order quantity is calculated as follows: 9 8 p 1 = < X Q ix max 0, Gi jIi1 ; :
j0

An example of MRP table for the POQ policy is given in Fig. 1. Each MRP table has several parameters: periodicity for grouping orders (for POQ policy), planned lead time for time phasing, safety stock (if necessary), etc. In this paper, only the following two essential parameters are considered: periodicity (p) and planned lead time (x) In Fig. 1, these parameters are equal to p 2 and x 1. In a stochastic environment, an adequate choice of these parameters is crucial, because this denes the average total cost of supply planning. In general, the calculation of the optimal values of these parameters is a very complex optimization problem. Note that in literature, another parameter is often studied, especially for a random demand, which is safety stock. This parameter is not considered here (safety stock is set to 0). Some discussions on safety lead time versus safety stock (the former is included in calculation of the planned lead time x) are reported in the next section.

2. Basic principles of MRP systems The goal of Material Requirements Planning (MRP) is to determine a replenishment schedule for a given time horizon. The MRP approach deals with the calculation of these requirements for a series of sequential planning periods (time buckets). One time bucket can be equal to a day, week or month depending on the applications. The gross requirements for the nished product are given by the Master Production Schedule (MPS). The net requirements and planned order releases are deduced from gross requirements and projected on-hand inventory. Let us introduce the following notations: x planned lead time, Q(i x) planned orders released at time bucket i x, and for time bucket i: I(i) projected on-hand inventory, N(i) net requirements, G(i) gross requirements. The initial inventory I(0) is assumed to be known. The net requirements for time bucket i are obtained as follows: Ni GiIi1 The planned released order quantity: Q ix maxf0, N ig This is the core of the MRP approach. These rules are implemented in MRP tables. The above formulas concern the Lot-for-Lot (L4L) policy, i.e. where the orders are not grouped. Often the simple L4L policy is

3. Related publications We have chosen to limit this section to some related publications and key articles illustrating the problem, its history and major results. For more exhaustive reviews on the simulation and analytical studies of MRP parameters under uncertainties, see Yeung et al. (1998), Koh and Saad (2003), Mula et al. (2006), Dolgui and Prodhon (2007), Inderfurth (2009) and Louly and Dolgui (2010). In general, the majority of publications are devoted to the MRP parameterization under customer demand uncertainties. Simulations are more often used. As to random lead times, the number of publications is relatively small. In this paper, the attention will be focused only on lead time uncertainties. For this case, some earlier publications have already considered a calculation of safety lead time via simulation. For example, the simulation experiments by Whybark and Wiliams (1976) show that under lead time uncertainty one is better offsetting safety times than safety stocks. Ho and Lau (1994) compared different lot-sizing rules. Molinder (1997) proposes a simulated annealing approach to nd appropriate safety lead time in a simulation/optimization approach. Enns (2001) showed that by selecting proper lot-sizes and planned lead times the inventory levels can be diminished. An alternative of MRP for the case of both demand and lead-time uncertainties was considered by Mohebbi et al. (2007). Using also simulation, the authors have shown that their approach, called capacity

Please cite this article as: Louly, M.-A., Dolgui, A., Optimal MRP parameters for a single item inventory with random replenishment lead time, POQ policy and service level.... International Journal of Production Economics (2011), doi:10.1016/j.ijpe.2011.02.009

M.-A. Louly, A. Dolgui / Int. J. Production Economics ] (]]]]) ]]] ]]]

driven procurement (CDP), could be appropriate in some cases of light to moderate production. As for analytical approaches, the literature is rich with stochastic models considering again a random demand of nished product for single item production systems. In this context, safety stocks can be determined by using, for example, different generalizations of the Newsboy model (Porteus, 1990; Petruzzi and Dada, 1999). Analytical models with stochastic lead times have been studied even less. Some better known historic examples are the paper of Kaplan (1970), which suggested a nite horizon dynamic programming approach, and Liberatore (1979), who tried to extend directly the EOQ model by introducing a stochastic lead time. Note that an interesting analytical evaluation of safety stock versus safety time in MRP environments was conducted by Buzacott and Shanthikumar (1994). Kumar (1989) examined an assembly system with stochastic component lead times and a xed production due date and quantity, as well as suggested a one-period model (for a sole batch). The timing of each order is determined so that the overall cost, composed of the holding and tardiness costs, is minimized. Fujiwara and Sedarage (1997) developed a model for a simple Just-in-Time (JIT) assembly system with random lead times. This is a continuous time model. The following general assumptions were used: product demand is constant and known, and the production capacity is innite. The authors have taken into account the inventory holding, shortage and setup costs. A decomposition algorithm is proposed where each sub-problem is solved numerically. More recently there has been published several other interesting articles on subjects related to our research. First of all, we would like to cite the paper of Hegedus and Hopp (2001), who proposed a practical method for setting safety lead time for purchased components in assembly systems. The case of random demand and stochastic lead times with xed order period (FOP) lot-sizing rule was considered. Silver and Zufferey (2005) studied an industrial case of a sawmill in Canada. This study was characterized by a xed demand and random lead time. The production of the sawmill also concerned the drying of logs and seasonal changes in transportation time. A tabu search heuristic was employed. In Arda and Hennet (2006), a problem with one retailer and several suppliers was considered as a queueing theoretical model. There is single type of product; the supplier lead time is exponentially distributed. The demand follows a Poisson process. The authors proposed a (S 1,S) inventory control policy and a rule for selecting the supplier of each inventory replenishment order. Chauhan et al. (2009) studied an assembly system. Component procurement times were continuous random variables and associated distribution functions are known in advance. The objective was to determine the ordering time (a real decision variable) for each component such as to minimize the sum of expected holding and backlogging costs. A decomposition approach and simulated annealing algorithm to solve this problem were proposed. Rossi et al. (2010) considered a (R, S) inventory control policy for single item with non-stationary stochastic demand. A constraint programming approach was proposed to optimize the parameters of this model. The authors demonstrated that if the lead time is stochastic, then policy parameters are considerably affected compared to the case when the lead time is deterministic or equal to zero. Finally, an excellent analysis of the problems of production planning under uncertainties was written by Graves (2011). What is very interesting in this chapter that it is based mainly on the personal observations of this MIT professor from production planning practices in various industrial situations. Note also that an important contribution to MRP theory was made in publica and his colleagues, see for example the paper tions of Grubbstrom et al., 2010) and references included. (Grubbstrom

To provide some background of our research, we summarize here our essential publications on this subject. Note that in contrast with the majority of previously cited publications in this section, in the papers listed below, the lead times are considered as discrete random variables and planned lead times are discrete decision variables. Particular cases of assembly systems with random component lead times and L4L lot-sizing policy were considered in Louly and Dolgui (2002) and Louly et al. (2008). In these publications, all components have identical properties, i.e. the same lead time probability distribution and identical unit holding cost. The articles (Louly and Dolgui, 2002, 2004) consider the case of the objective function minimizing the sum of average holding and backlogging costs, while Louly et al. (2008) studies the case when backlogging cost is replaced with a service level constraint. For the latter case, in Louly and Dolgui (2004), the Periodic Order Quantity (POQ) policy was modeled and some properties of the objective function were proven. These properties were used in Louly and Dolgui (2010) to develop a Branch and Bound algorithm. In this article, using the results of Louly and Dolgui (2004, 2010), obtained for component procurements for assembly systems, we propose an optimal formulation for the case of single item replenishments. Note that this formulation is a new extension of the well-known Newsboy model. Let us summarize this state of the art. First of all, most of existing models that take into account random lead times deal with particular cases with continuous decision variables, specic inventory policies and normal or exponential probability distributions of lead times. Simulation, heuristic techniques or decomposition approaches are used to solve these models. For the cases when discrete decision variables are employed, usually enumeration approaches, such as Branch and Bound, are developed. In contrast, in this paper, we propose a general analytical solution for any discrete probability distributions of lead time. Of course, to obtain this solution, we limited our study to only a single item inventory with a xed demand. Nevertheless, the model obtained is quite general and can be used directly for MRP parameterization or as an element of heuristic for more complex procurement situations.

4. Optimization problem 4.1. Inventory model To take into account the particularities of MRP parameterization, the following model will be considered in this paper:

 Items are ordered from external suppliers to satisfy the

customer demand.

 Probability distribution of procurement time (lead time) L is

known. There is an upper value of the lead time distribution, i.e. L r u. The customer demand is known and is constant for all time buckets and equal to D. The unit holding cost h per period, and setup cost c are also known. POQ policy is used: items are ordered at every p periods. The orders for items are made at the beginning of the periods kp + 1, k 0,1,2,y, and there is no order made in the periods kp + r, r 2,3,y, p. Then, the supply orders Q are constant Q Dp (p is a decision variable). The nished product demands are satised at the end of each period and unsatised demands are backordered and have to be satised during the subsequent time buckets. The goal of this model is to search for the optimal values of the parameters p and x.

 

Please cite this article as: Louly, M.-A., Dolgui, A., Optimal MRP parameters for a single item inventory with random replenishment lead time, POQ policy and service level.... International Journal of Production Economics (2011), doi:10.1016/j.ijpe.2011.02.009

M.-A. Louly, A. Dolgui / Int. J. Production Economics ] (]]]]) ]]] ]]]

The following notations are used: 1e h c u Lk D p Q x Z+ desired service level unit holding cost setup cost, i.e. the cost of a supply order upper value of lead time distribution lead time of items ordered at the beginning of the period k demand per period periodicity of replenishments supply order quantity planned lead time max(Z,0)

cumulative required quantity D(kp +r) minus the initial item inventory Dx minus the delivered quantity Dp(k+ 1 Np,kp + r). Then, the number of backlogged demands is equal to D(pNp,kp + r + r p x) + , while the inventory is its inverse. Thus, the total satised demand is equal to D[kp + r (pNp,kp + r + r p x)]. Given that there is a setup cost c if r 1, and using Z + Z +Z , the sum of the holding and setup costs at the end of period m can be rewritten as c 1r 1 Dhx pr pN p, kp r DhpN p, kp r r px As shown in the previous proposition, the cost of a single period kp +r is a random variable. To study the considered multiperiod problem, explicit closed forms should be obtained for the average cost and the average number of shortages on the innite horizon, i.e. for the following expressions: C X , p lim
m 1 X k1

As the lead time is a random variable, a planned lead time can be greater than the corresponding forecasted (or contracted) lead time. The difference is a safety lead time. In the model considered, the demand D is constant and the quantities ordered are the same and equal to Dp; so the optimal planned lead time x is also the same for all orders; thus it provides also initial inventory Dx (x/p orders of Dp items). Therefore, the aim of this study can be expressed in other terms: the goal is to nd the optimal values of the parameter p and initial inventories Dx, where x is the planned lead time. In this paper, an approach is proposed to optimize the planned lead time x and the periodicity p of POQ policy minimizing the sum of setup and holding costs while respecting a service level constraint. The method suggested takes into account the fact that the actual lead times are random. 4.2. Analytical expression of the criterion

m-1 m

C X , p, N p, k

SX , p lim

m 1 X 1SX , p, Np,k 4 0 m-1 m k1

Theorem 1. The average cost and average number of shortages have the following closed forms: C X , p 1X EC X , p, Np, kp r pr1
 ! p X c p1 1 X p, r x kr p h hxEN p h 1 F p 2 pr1 p kZ0
p

Some proofs will be made here to obtain an analytical expression for the criterion as a function of decision variables. For the considered model, given that the maximal value of the lead time is equal to u, only the orders made in the earlier u 1 periods may not have arrived yet. The orders made before have already arrived. This property can be used to obtain the distribution of probability for the number Np,m of expected deliveries at the end of the period m kp + r. Let Lm + 1 j, j r, r + p, r +2p,y, r + ((u 1 r)/p)p be the lead time of the orders made at the beginning of the periods kp + 1, (k 1)p + 1,y, (k ((u 1 r)/p))p +1. If Lm + 1 j 4 j, then the order made in the period m + 1 j is delivered after the end of the period m. Let 1E be the binary function equal to 1 when the expression E is true, otherwise it is equal to 0. So, if 1Lm 1j 4 j is equal to 1, then the order made at the period m + 1 j is delivered after the end of the period m. Thus, the random variables Np,m can be represented as follows: Np, m Np, kp r r A f1, 2, . . ., pg
u 1r =p X j0

6 SX , p 1X PrpN p, kp r r px 4 0 pr1   p 1 X p, r xr p 1 F pr1 p

p

where F p, r x PrNp, r r x PrN p, kp r r x. Proof. For each value of p within the set {1, 2, y, u 1}, the following p stochastic processes are studied: N p, kp r Pu1r=p 1Lkjp 1 4 jp r , rA{1,2,..,p}, k Z 0. j0 For each value of r, the sequence of random variables Np,kp + r, k Z 0 constitutes a discrete stochastic process. The Law of Large Numbers cannot be applied here, because of the dependence among these variables; this is why, this proof will be based on a covariance stationary process. In fact, it is enough to prove that (i) the functions are measurable (ii) Np,kp + r, k Z 0 is a covariance stationary process and (iii) lim CovNp, kp r , N p, k tp r 0
t -1

1Lkjp 1 4 jp r , k Z 0, p A f1, 2, . . ., u1g, 1

The variables Np,m are independent from the decision variable x. Thus, they can be used to derive closed forms for the cost function (Louly and Dolgui, 2004). Proposition. Cost for the period m kp + r, k Z 0, pA{ 1,2,..,u 1 }, rA{ 1,2,..,p } is equal to C x, p, Np, kp r c 1r 1 Dhx pr pN p, kp r DhpN p, kp r r px 2 Proof. To obtain an explicit expression for the cost function it is sufcient to derive the expression for Np,m. In fact, the number of backlogged demands at the end of period m is equal to the

Condition (i) is usually veried because the criterion function is composed of measurable functions. Condition (ii) is veried because the lead times for the orders made at different periods are iid random variables. Then, the distribution of Np,kp + r does not depend on k. So, {Nk,k Z 1} is a stationary process. This condition is stronger than (ii); so (ii) is also valid. Finally, condition (iii) is veried due to the fact that Np,kp + r and p,(k + t)p + r N are independent, for t Z u, thus, lim CovNk , Nk t 0.
t-1

Please cite this article as: Louly, M.-A., Dolgui, A., Optimal MRP parameters for a single item inventory with random replenishment lead time, POQ policy and service level.... International Journal of Production Economics (2011), doi:10.1016/j.ijpe.2011.02.009

M.-A. Louly, A. Dolgui / Int. J. Production Economics ] (]]]]) ]]] ]]]

Therefore, the cost can be written as follows: C X , p 1 EC X , p, Np, kp r pr1

p X

Table 1 Probability distributions for N3,r. N3,r 0 r1 r2 r3 0.0280 0.0802 0.1623 N3,r 1 0.3060 0.4195 0.5027 N3,r 2 0.5032 0.4204 0.3070 N3,r 3 0.1628 0.0800 0.0280

By calculating the expectation and probabilities, we obtain the nal closed forms as follows:  ! p X c p1 1 X p, r x kr p p C X , p 1 F h hxEN h p 2 pr1 p kZ0   p 1 X p, r xr p SX , p 1 F pr1 p 4.3. Optimal solution properties The optimization problem can be written as follows: Min C X , p 8

Table 2 Rao-Cramer bounds for probability distributions of N3,r. r1 r2 r3 0.0000000272 0.0000000738 0.0000001360 0.0000002124 0.0000002435 0.0000002500 0.0000002500 0.0000002437 0.0000002128 0.0000001363 0.0000000736 0.0000000272

Table 3 Optimal planned lead time and corresponding cost for each periodicity. p x C(x,p) 1 7 12.50 2 8 9.01 3 8 7.84 4 9 8.50 5 9 8.49 6 9 8.66 7 9 8.93 8 9 9.25 9 9 9.60

Subject to   1 xr p Z 1e F p, r pr1 p
p X

9 10 11 12 13

F p, r x PrNp, r r x 0 r x r u1 Np
p X r1

N p, r

1 r p r u1

The optimal planned lead time corresponding to each periodicity p can be calculated using the following theorem. Theorem 2. For each given periodicity p, the optimal planned lead time x is the smallest integer, satisfying the following inequality:   p 1X xr p Z 1e Pr N p, r r 14 pr1 p Proof. We introduce a function G(x,p) as follows: Gx, p C x 1, pC x, p  ! p X 1 X p, r x k pr 1 F hh pr1 p kZ0  !! p X X 1 x 1 k pr 1 F p, r pr1 p kZ0  !   p p X X 1 x pr p, r x pr hh 1 F p, r F h p pr1 p p r1

random variables with parameters PrLj 4 jp r , j 0, 1, . . ., (u 1 r)/p+ 1. To illustrate the quality of this approximation, we present the probability distributions obtained for r 1,2,3, and p 3 in Table 1 and the corresponding Rao-Cramer bounds in Table 2. After that, the optimal lead times x are obtained with Matlab by using (6) and (14), for each periodicity p. The results are reported in Table 3. The global optimal solution is obtained when p 3 and x 8. Its cost C(x,p) is 7.84. Note that for L4L policy (p 1) the cost is 12.5. Thus, a substantial cost saving is possible when we use POQ policy instead of L4L. Note also that in this example optimal values of x are not always the same for different values of p. Nevertheless, the values of x change slowly over p. Therefore, the most important factor for x is the objective service level. Indeed, Theorem 2 shows that x does not depend on costs c and h. These two costs have, however, a signicant inuence on the parameter p.

5. Concluding remarks The obtained model is useful for the optimization of the MRP parameterization with random procurement times and the POQ lot-sizing policy. The proposed model minimizes the sum of the average holding and setup costs while satisfying a constraint on a desired service level. This is a new generalization of the Newsboy model. It can be used in many industrial situations. For example, often security coefcients are introduced to calculate the planned lead time for unreliable suppliers in an MRP environment. In this case, planned lead time is equal to contractual (or forecasted) lead time multiplied by a security coefcient. This coefcient is empiric but anticipates the delay by creating safety lead time. The more unreliable a supplier is, the larger its coefcient. The model suggested in this paper can be used to better estimate these coefcients by using statistics on the procurement lead times for each supplier and taking into account the holding and setup costs. This is a multi-period model with no major restriction on the type of the lead time distribution. All discrete distributions can be used. The decision variables are integer; they represent the periodicity and planned lead time for items. Concerning the assumption of constant demand, note that this model should be used with different possible values of the

15

On the other hand, Fp,r are positive functions, due to constraint (10). Then, G(x,p) is a positive function. This means that C(x,p) is an increasing function on x. The optimal planned lead time will then be the smallest integer satisfying all the constraints. Fp,r are cumulative functions. Using their denition given by (10) in the service level constraint (9), we obtain the inequality (14). Let us give a numerical example. Example. c 10; h 1; u 10. The lead time is uniformly distributed between 1 and 10. We solved this problem for the case when the desired service level is equal to 99%. First, the probability distributions for Np,r were approximated by using (1). In fact, for each pA { 1,2,y,u 1 } and rA{ 1,2,y,p }, the value Np,r can be seen as a sum of independent Bernoulli

Please cite this article as: Louly, M.-A., Dolgui, A., Optimal MRP parameters for a single item inventory with random replenishment lead time, POQ policy and service level.... International Journal of Production Economics (2011), doi:10.1016/j.ijpe.2011.02.009

M.-A. Louly, A. Dolgui / Int. J. Production Economics ] (]]]]) ]]] ]]]

demand to examine the sensitivity of the obtained parameters to the said values. If the parameter values are signicantly different for the given demand levels, the approach by scenarios can be applied to choose the best parameter values. In addition, the demand variations can be decoupled from planned lead time calculation by using safety stocks. This is another promising perspective for future research.

Acknowledgments The authors thank Chris Yukna for his help in English. References
Arda, Y., Hennet, J.C., 2006. Inventory control in a multi-supplier system. International Journal of Production Economics 104 (2), 249259. Axsater, S., 2006. Inventory Control, 2nd ed Springer. Baker, K.R., 1993. Requirement planning. In: Graves, S.C. (Ed.), Handbooks in Operations Research and Management Science, Logistics of Production and Inventory, vol. 4. North-Holland, Amsterdam, pp. 571628. Buzacott, J.A., Shanthikumar, J.G., 1994. Safety stock versus safety lead time in MRP controlled production systems. Management Science 40 (12), 16781689. Chauhan, S.S., Dolgui, A., Proth, J.-M., 2009. A continuous model for supply planning of assembly systems with stochastic component procurement times. International Journal of Production Economics 120 (2), 411417. Chaharsooghi, S.K., Heydari, J., 2010. LT variance or LT mean reduction in supply chain management: which one has a higher impact on SC performance? International Journal of Production Economics 125, 475481. Dolgui, A., Prodhon, C., 2007. Supply planning under uncertainties in MRP environments: a state of the art. Annual Reviews in Control 31 (2), 269279. Dolgui, A., Proth, J.-M., 2010. Supply Chain Engineering: Useful Methods and Techniques. Springer. Enns, S.T., 2001. MRP performance effects due to lot size and planned lead times setting. International Journal of Production Research 39 (3), 461480. Fujiwara, O., Sedarage, D., 1997. An optimal (Q,r) policy for a multipart assembly system under stochastic part procurement lead times. European Journal of Operational Research 100, 550556. Graves, S.C., 2011. Uncertainty and production planning. In: Kempf, K.G., Keskinocak, P., Uzsoy, R. (Eds.), Planning Production and Inventories in the Extended Enterprise. A State of the Art Handbook. Series: International Series in Operations Research & Management Science, vol. 151, pp. 83102. Grubbstrom, R.W., Bogataj, M., Bogataj, L., 2010. Optimal lotsizing within MRP Theory. Annual Reviews in Control 34 (1), 89100. Hegedus, M.G., Hopp, W.J., 2001. Setting procurement safety leat-times for assembly systems. International Journal of production Research 39 (15), 34593478. Ho, C.-J., Lau, H.-S., 1994. Evaluating the impact of lead time uncertainty in material requirements planning systems. European Journal of Operational Research 75, 8999. Inderfurth, K., 2009. How to protect against demand and yield risk in MRP systems. International Journal of Production Economics 121, 474481.

Kaplan, R.S., 1970. A dynamic inventory model with stochastic lead times. Management Science 16, 491507. Koh, S.C.L., Saad, S.M., 2003. MRP-controlled manufacturing environment disturbed by uncertainty. Robotics and Computer Integrated Manufacturing 19, 157171. Kumar, A., 1989. Component inventory cost in an assembly problem with uncertain supplier lead-times. IIE Transactions 21 (2), 112121. Lee, H.L., Nahmias, S., 1993. Single-product, single-location models. In: Graves, S.C. (Ed.), Handbooks in Operations Research and Management Science, Logistics of Production and Inventory, vol. 4. North-Holland, Amsterdam, pp. 357. Liberatore, M.J., 1979. The EOQ model under stochastic lead times. Operations Research 27, 391396. Louly, M.A., Dolgui, A., 2002. Newsboy model for supply planning of assembly systems. International Journal of Production Research 40, 44014414. Louly, M.A., Dolgui, A., 2004. The MPS parameterization under lead time uncertainty. International Journal of Production Economics 90, 369376. Louly, M.A., Dolgui, A., 2010. Optimal time phasing and periodicity for MRP with POQ policy. International Journal of Production Economics, doi:10.1016/j.ijpe. 2010.04.042. Louly, M.A., Dolgui, A., Hnaien, F., 2008. Supply planning for single-level assembly system with stochastic component delivery times and service-level constraint. International Journal of Production Economics 115 (1), 236247. Mohebbi, E., Choobineh, F., Pattanayak, A., 2007. Capacity-driven vs. demanddriven material procurement systems. International Journal of Production Economics 107, 451466. Molinder, A., 1997. Joint optimisation of lot-sizes, safety stocks and safety lead times in an MRP system. International Journal of Production Research 35, 983994. Mula, J., Poler, R., Garca-Sabater, J.P., Lario, F.C., 2006. Models for production planning under uncertainty: a review. International Journal of Production Economics 103, 271285. Mula, J., Poler, R., 2010. Fuzzy material requirement planning. In: Kahraman, C., Yavuz, M. (Eds.), Production Engineering and Management under Fuzziness, vol. 252. Springer, pp. 3957 Studies in Fuzziness and Soft Computing. Petruzzi, N.C., Dada, M., 1999. Pricing and the newsvendor problem: a review with extensions. Operations Research 47, 183194. Porteus, E.L., 1990. Stochastic inventory theory. In: Heyman, D.P., Sobel, M.J. (Eds.), Hand-book in OR and MS, vol. 2. Elsevier Science Publishers, pp. 605652. Rossi, R., Tarim, A.S., Hnich, B., Prestwich, S., 2010. Computing the non-stationary replenishment cycle inventory policy under stochastic supplier lead-times. International Journal of Production Economics 127, 180189. Silver, E.A., Zufferey, N., 2005. Inventory control of raw materials under stochastic and seasonal lead times. International Journal of Production Research 43, 51615179. Sipper, D., Buln, Jr, R.L., 1998. Production: Planning, Control and Integration. McGraw Hill. Tempelmeier H., 2006. Inventory Management in Supply Networks: Problems, Models, Solutions, Books on demand GmbH. Whybark, D.C., Wiliams, J.G., 1976. Material requirement under uncertainty. Decision Science 7, 595606. Yeung, J.H.Y., Wong, W.C.K., Ma, L., 1998. Parameters affecting the effectiveness of MRP systems: a review. International Journal of Production Research 36, 313331. Zipkin, P., 2000. Foundation of Inventory Management. McGraw-Hill.

Please cite this article as: Louly, M.-A., Dolgui, A., Optimal MRP parameters for a single item inventory with random replenishment lead time, POQ policy and service level.... International Journal of Production Economics (2011), doi:10.1016/j.ijpe.2011.02.009