Priority optimization and make‐to‐stock/make‐to‐order decision in multiproduct manufacturing systems

• Published date: 01 July 2018
• Author: K. Hadj Youssef,Ch. van Delft,Y. Dallery
• DOI: http://doi.org/10.1111/itor.12464
• Date: 01 July 2018

Intl. Trans. in Op. Res. 25 (2018) 1199–1219

DOI: 10.1111/itor.12464

INTERNATIONAL

TRANSACTIONS

IN OPERATIONAL

RESEARCH

Priority optimization and make-to-stock/make-to-order

decision in multiproduct manufacturing systems

K. Hadj Youssefa, Ch. van Delftband Y. Dalleryc

aLaboratoire de G´

enie M´

ecanique, Ecole Nationale d’Ing´

enieur de Monastir, 5000 rue Ibn El Jazzar, 5035 Monastir,

Tunisie

bInformation Systems and Operations Management Department, GroupeHEC Paris, 1 route de la lib´

eration, 78351

Jouy-en-Josas, France

cLaboratoire de G´

enie Industriel, Ecole Centrale Paris, GrandeVoie des Vignes, 92290 Chˆ

atenay-Malabry, France

E-mail: Khaled.HajYoussef@enim.rnu.tn [Hadj Youssef]; vandelft@hec.fr [van Delft]; dallery@lgi.ecp.fr [Dallery]

Received 5 July2016; accepted 15 August 2017

Abstract

We consider a single-stage multiproduct manufacturing facility producing a large number of end products.

In order to reduce overall inventory costs, an efficient approach is to produce some items according to a

make-to-stock (MTS) policy and others according to a make-to-order (MTO) policy. Items priority levels

play a key role in the optimal MTO/MTS decisions for such typical large-scale systems. To tackle this issue,

the manufacturing facility is modeled as a multiproduct multipriority classes queuing system. We propose

a general optimization procedure that selects near-optimal priority classes, gives the associated flow control

mode (MTO or MTS) for each product, and provides a lower bound and an upper bound with respect to the

optimal cost. First, we illustrate efficiency of our optimization procedure for this class of nonlinear integer

programs via several examples and by a numerical analysis, including a comparison with two alternative

heuristics given in the literature. In addition, we provide managerial insights by exhibiting, under various

parameter settings, the significant impact of an efficientpriority level allocation among items on the inventory

costs and on optimal splitting between MTO and MTS products.

Keywords: make-to-stock (MTS); make-to-order (MTO); priority level; heterogeneous multiproduct produc-

tion/inventory system; queuing model

1. Introduction

An important tactical issue in production management is whether products should be manufac-

tured according to a make-to-order (MTO) or a make-to-stock (MTS) policy. By definition, under

MTO management, a production order is released to the manufacturing facility only after a firm

C

2017 The Authors.

International Transactionsin Operational Research C

2017 International Federation of OperationalResearch Societies

Published by John Wiley & Sons Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main St, Malden, MA02148,

USA.

1200 K. Hadj Youssef et al. / Intl. Trans. in Op. Res.25 (2018) 1199–1219

order has been received from the customer. On the contrary, under MTS management, products

are manufactured in anticipation of future demands. They are stored in the finished goods inven-

tory from which the customer orders will be directly delivered. Adopting a pure MTS policy or

a pure MTO policy for all the products can be considered as extreme choices as there exists the

option of combining these two policies. As a typical example, the PC’s assembly industry keeps

available in distributors inventories some basic (standard) products, while the products ordered

with specific options defined by the customer are delivered in MTO within a specified lead time.

Other examples can be cited such as food industry (see Soman et al., 2007), semiconductor industry

(see Wu et al., 2008), or steel industry (see Tao et al., 2015). In such a combined setting, efficient

MTO/MTS allocation is well known to depend on each product-type parameters, such as the

demand behavior (intermittent vs. regular, low volume vs. high volume), required customer lead-

time (long vs. short), or inventory holding cost (high vs. low). But, in addition to the product pa-

rameters, feasibility of MTO/MTS decision, as well as associated inventory cost performance, also

depends on congestion effects induced by the demand processes and by the priority mechanisms

among the production orders. In order to intuitively illustrate the importance of priority effects,

let us start with a production/inventory setting where there is no specific priority rule between

the work orders. In this case, congestion mechanisms effectively act similarly on all production

orders, and items with very short admissible delivery lead-time automatically have to be managed

in MTS because MTO is not feasible for them. If, on the contrary, some items are prioritized,

the congestion effect for these items is reduced, which could permit one to produce them under

an MTO policy, while still satisfying the delivery lead-time required by the customers. Clearly, a

trade-off appears as the congestion effect for the nonprioritized items will increase and additional

inventories will be necessary for these items. The key point is that in typical large-scale prob-

lems, when the priority allocation is optimized, the trade-off can be expected to be significantly

positive.

The purpose of this paper is to provide a general optimization procedure that gives op-

timal flow control (MTO or MTS) to associate to each product when taking into ac-

count congestion effects and inventory costs associated with an optimized priority allocation

strategy.

In the tactical MTO/MTS decision analyzed here, the optimization of the priority mecha-

nism does not aim at solving the real-time scheduling policy at the operational level, but rather

to capture the main global effects linking priority, congestion, and MTO/MTS optimization.

Once this MTO/MTS decision has been made, the operational flow management decisions at

the shop floor level will determine the best short-term scheduling decisions within the frame-

work specified at the tactical level. Examples of such short-term scheduling policies can be found

in Chang et al. (2003), Kaminsky and Kaya (2009), Soman et al. (2006, 2007), and Wu et al.

(2008).

Many research studies have addressed the performance analysis and optimization of combined

MTS and MTO production/inventory systems (for a review, see Soman et al., 2004). Several papers

have analyzed the problem as a tactical planning issue on a finite horizon via discrete-time models

with capacity constraints (see Choi, 2014; Khakdaman et al., 2015; Tao et al., 2015). However,

the congestion effects associated to capacity sharing among products are not taken into account

by these authors. Some other works studied the optimality of MTO versus MTS policies for a

multiproduct manufacturing system where the products are scheduled according to a FIFO policy

C

2017 The Authors.

International Transactionsin Operational Research C

2017 International Federation of OperationalResearch Societies