Introduction of a State Feedback Structure for Adjusting Priority Rules in Production Scheduling

Abstract

This paper proposes a method of generating and selecting rules for adjusting the priorities of jobs by using a genetics-based machine learning (GBML) technique, where a state feedback structure is newly introduced. In applying the GBML, we use the Pitts approaches, where the set of rules (rule-set) are represented symbolically as an individual of genetic algorithms, and the fitness of an individual is calculated based on the makespan of the schedule generated by using the rule-set. As for a rule representation, we consider several attributes and status of a job in order to calculate its priority as a weighted sum of these attributes. We actually carried out computational experiments to simple problems with an intree-type precedence relation, which indicate that the method of applying the GBML is effective for finding and selecting good rule-sets.