Capacity decision under preset service level and process flexibility: A combined application of genetic algorithm and sample-based optimization

Abstract

Purpose

Process flexibility (PF) is seen as a hedging instrument against demand uncertainty. This paper aims to examine capacity decisions for both flexible and dedicated processes under production policies such as make-to-order and make-to-stock. The study identifies some relative benefits, in terms of expected profit, of the process flexible plant over the dedicated ones. Furthermore, the advantage appears to be contingent upon the decision on the preset service level.

Design/methodology/approach

Using the sample-based optimization procedure, a detailed computational analysis is undertaken to identify the conditions under which a flexible plant is preferred over a dedicated plant. A combination of genetic algorithm and sample-based optimization procedure is used to capture the effects of preset service level. The factors controlled in this paper include the demand variance, demand correlation, capacity investment cost and the product price.

Findings

According to this study, in a dedicated process changing to a flexible process is not justified for the same level of demand correlation even with high demand variance. In fact, a strict control on the preset service level prefers the dedicated strategy. The advantage of a flexible plant increases as the demand correlation decreases, product price decreases, price asymmetry increases or capacity investment cost increases. With a preset service level constraint, a flexible process should be preferred to a dedicated one only when the capacity investment cost is high or the products have low contribution margins.

Originality/value

The PF index is introduced in this paper to measure the benefit of a flexible plant over a group of dedicated plants. The benefits were found to be contingent upon the decision on the required service level.

Citation

Bose, D., Chatterjee, A.K. and Barman, S. (2018), "Capacity decision under preset service level and process flexibility: A combined application of genetic algorithm and sample-based optimization", Journal of Modelling in Management, Vol. 13 No. 3, pp. 755-772. https://doi.org/10.1108/JM2-08-2017-0086