Adopting Due-Date Buffer for Real-Time Scheduling Using Lagrangean Relaxation Method in an APS Environment

摘要

Real-time Scheduling has been recognized as the central function in an Advanced Planning and Scheduling (APS) environment. The present study proposes a hybrid approach of utilizing the Lagrangean Relaxation method and due-date buffer in order to realize the real-time scheduling function. A penalty cost is calculated from a relevant due-date buffer is incorporated into the normal Lagrangean Relaxation method by two approaches. The implicit approach is to adopt the penalty cost in the dispatching rule of the list scheduling, and the explicit approach is to adopt the penalty cost in the objective function of the Lagrangean Relaxation method. The proposed approaches are compared to the original methods by various experiments adopting different features and scales of the workshops. Based on the experiment results, the effectiveness of the proposed approaches is verified and a feasible experimental method by which to decide the best due-date buffer size is presented.