Seru scheduling problem with lot streaming and worker transfers: A multi-objective approach

Abstract

Seru production system (SPS) offers the flexibility of job shop production environments with the efficiency of traditional assembly lines. The SPSs are particularly attractive to industries characterized by high product variety and micro production volumes, and effective utilization of production and workforce resources is a critical challenge for SPSs. This paper addresses the seru scheduling problem with lot streaming and worker transfers for a SPS using a multi-objective approach. To this end, first, a multi-objective mixed-integer linear programming (MILP) model is developed for the minimization of makespan, average flow time, and maximum workload imbalance. Six different algorithms based on non-dominating sorting genetic algorithm II (NSGA-II) are developed, each corresponding to an operational setting dictated by the lot streaming and worker transfers strategies in effect. A design of experiment (DoE) framework is utilized to generate realistic problem instances based on the several controllable factors and their levels. Analysis of comprehensive computational results demonstrates the effectiveness of the proposed algorithm (NS2) in finding high-quality and diversified solutions by simultaneous utilization of lot streaming with variable-sized sublots and worker transfers. The results indicate that the performance improvement achieved by the NS2 ranges between 10% and 20% compared to other algorithms. Furthermore, Analysis of Variance (ANOVA) confirms the significance of the number of workers and number of serus as critical parameters for the design or redesign of SPSs. Drawing on these findings, managerial insights are provided regarding the impact of lot streaming and worker transfers on SPS performance. This study offers practical and theoretical insights for decision-makers seeking to enhance SPS performance and bridge the gap between the conceptual analysis and practical implementation of SPSs.