A bit-wise mutation algorithm for mixed-model sequencing in JIT production systems

Abstract

Product sequencing in a mixed-model production system is an operational level problem and is computationally hard. This necessitates the use of non-traditional optimisation methods to solve it in a short time. In the first phase, this study addresses the mixed-model sequencing problem with the objective of minimising production rates variation (PRV) which is a just-in-time objective. We formulated a mixed integer quadratic assignment model for this problem. LINGO 14.0 solver could solve this model for small- and medium-size problems. The study presents a bit-wise mutation algorithm to solve the sequencing problem with the same objective. This algorithm uses a single unary operator which makes it simpler and faster. Computational results show that the algorithm succeeds in solving large-size PRV problems in a reasonable time, for which the algorithms reported in the literature could only get near-optimal solutions. In the second phase, we consider both PRV and one of the system performance measures, makespan as objectives for the mixed-model sequencing problem. A mathematical description of the underlying model is provided. The bit-wise mutation algorithm is modified to generate a set of non-dominated solutions which provide the decision-maker with the opportunity to trade-off between the two objectives.

Keywords:

• JIT
• mixed-model production sequencing
• production rates variation (PRV)
• bit-wise mutation algorithm
• makespan
• non-dominated solutions

View correction statement:

Corrigendum

Acknowledgements

The authors would like to thank the two anonymous reviewers for their valuable comments and suggestions, which were useful for improving the study, considerably.

Disclosure statement

No potential conflict of interest was reported by the authors.