Dispatching rules for dynamic scheduling of one-of-a-kind production

Abstract

This paper presents an extensive performance analysis of dispatching rules for dynamic scheduling in one-of-a-kind production (OKP) where each job has its own due-date and multilevel routing structure. For the analysis, a framework of dynamic dispatching rules is proposed to identify five dispatching rules: job due-date (JDD) rule, normalized JDD rule, latest finish date (LFD) rule, latest start date (LSD) rule, and refined LSD rule. The performance analysis shows that: (i) JDD, LFD and LSD rules perform fairly well, which is compatible with the results of existing studies; (ii) the refined LSD rule outperforms other rules; (iii) the normalized JDD rule performs very poorly. Also presented is a framework of simulation-based dynamic scheduling in which the refined LSD as well as the 'try-every-rule' approach can be effectively utilized.