Cycle scheduling for discrete shipping and dynamic demands

Abstract

This study examines previously unexplored issues regarding the performance of integrative cycle scheduling ((ICS), i.e., lot sizing and sequencing) in a multi-item, capacitated, repetitive manufacturing environment with discrete shipping and dynamic demands. Through an extensive experimental study, we provide a more complete understanding and managerial insights on cycle scheduling for discrete shipping and dynamic demands. Experiment 1 identifies the dominance relationship between lot sizing and sequencing in the cycle scheduling performance in the repetitive manufacturing environment with different daily demand variations, by comparing a large set of ICS and other heuristics in terms of total inventory costs and CPU time. Experiment 2 investigates the performance sensitivity of the ICS heuristics under different capacity utilization factor settings determined by capacity tightness, setup time, and processing time. And experiment 3 examines the robustness of the ICS in initial inventory buildup for feasibility and rolling horizon scheduling situations. Important managerial insights into cycle scheduling from the findings, besides the overall favorable performance of the ICS heuristics, include: (1) the situation-dependent dominance of different lot sizing rules, i.e., the best cost performance of the modified Pinto–Mabert based heuristics for the low demand variation and the common cycle based heuristics for the high variations, and the good cost performance of selecting the best feasible sequence from random sampling, (2) the crucial role of capacity factors in procedure performance, and (3) the robust initial inventory buildup performance of the ICS in a changing demand environment.