Multi objective lotsizing and scheduling with material constraints in flexible parallel lines using a Pareto based guided artificial bee colony algorithm

Abstract

Flexible production lines are significantly used in recent years for production of different variety of products in large quantity. These production lines usually produce mixed model products in lots which can improve the production efficiency and satisfy the different customer’s demand. However, it is very challenging for factory planners to make decision on the optimal size of lots for each product model in production lines. Therefore, current research investigates lotsizing and mixed model scheduling problem in flexible parallel production lines considering sequence dependent setup times between mixed model products with aims to minimize makespan of production lines, balance the workload among lines and maximize the net profit, simultaneously. Additionally, a new constraint of material availability is introduced to the problem. A novel Pareto based guided artificial bee colony algorithm (PGABC) with three different guided mechanisms is designed for the current problem to obtain near optimal Pareto solutions. Taguchi method is adopted to tune the effective parameters of the proposed PGABC algorithm. Furthermore, nine different sets of instances including 90 problems are generated and tested using PGABC. The performance of PGABC is compared with other three famous methods used for multi objective optimization in literature including, multi objective artificial bee colony algorithm, non-dominated sorting genetic algorithm III (NSGAIII) and improved strength Pareto evolutionary algorithm (SPEA2). Computational results indicate that proposed PGABC outperforms the other considered algorithms both in terms of solution diversity and quality based on the considered test problem instances. Finally, an industrial case problem from a bus and coach company is solved by the proposed approach for mixed-model lotsizing and scheduling in parallel production lines.