Predicting the performance of rescheduling strategies for parallel machine systems

Abstract In dynamic, stochastic manufacturing systems, production planners and manufacturing engineers can benefit from understanding how rescheduling strategies affect system performance. This knowledge will help these experts design and operate better manufacturing planning and control systems. This paper presents new analytical models that can predict the performance of rescheduling strategies and quantify the trade-offs between different performance measures. In the parallel machine systems under consideration, jobs of different types arrive dynamically, and setups occur when production changes from one job type to another. Three rescheduling strategies are studied: periodic, hybrid, and event-driven based on the queue size. The scheduling algorithm groups jobs of the same type in batches to eliminate unnecessary setups. The analytical models require less computational effort than simulation models, and experimental results show that they accurately estimate important performance measures like average flow time, machine utilization, and setup frequency.