A multi-agent model for reactive job shop scheduling

Abstract

Scheduling can be divided into two classes, namely, predictive and reactive scheduling. The former addresses problems with initially available jobs, deterministic processing times and available machines throughout the scheduling horizon; the latter has problems with random job arrivals, non-deterministic processing times and unpredictable events such as machine breakdowns. This paper presents a complete multiple agents' framework for reactive job shop scheduling. It provides a theoretical justification that this approach is actually a completely reactive scheduling approach combining real time decision making with predictive decision making and can resolve various disruptions as flexibly as dispatching rules. The justification also includes solution schemes for dynamic job arrivals, which make full use of available information for further scheduling robustness enhancement.