JIT scheduling rules: A simulation evaluation

Abstract

Just-In-Time (JIT) production systems capitalize on simplicity and the ability of workers to make decisions in a decentralized manner. In a multiproduct line operating under kanban control, the production worker at a given station must schedule the different jobs awaiting processing using information available locally. In practice, the first-come-first-served (FCFS) rule is commonly used. Recent results reported in the literature indicated that the shortest-processing-time (SPT) rule performed better than the FCFS rule. In this paper, we provide a simulation evaluation of the performance of a number of scheduling rules operating under different JIT production scenarios. Our hypothesis is that there are differences in the relative performance of the scheduling rules under different production scenarios. We differentiate among the JIT scenarios by the extent of setup time reduction already carried out (as indicated by the ratio of setup to processing times), the amount of slack in the system (as measured by the number of kanbans circulating), the extent to which uncertainty has been eliminated (as determined by the stochasticity of processing times), and the complexity of production requirements (as specified by the product-mix in mixed-model assembly). In this way, this paper provides further insights into the performance of scheduling rules operating under different JIT production environments, thereby adding to the scope and depth of research in this particular aspect of JIT production systems.