Category: Evolutionary Scheduling Learning to Improve Genetic Algorithm Performance on Job Shop Scheduling

This paper presents a new approach to genetic algorithm based scheduling. We use genetic algorithms augmented with a case-based memory, containing population members from past problem solving attempts, to obtain better performance over time on sets of similar job shop scheduling problems. Rather than starting with a randomly initialized population on each scheduling problem, we periodically inject a genetic algorithm's population with appropriate cases (encoded schedules) from similar, previously solved problems. Experimental results on randomly generated sets of 10x10 and 15x15 job shop scheduling problem demonstrates the performance gains from our approach and show that our system learns to take less time to provide quality solutions to new scheduling problems as it gains experience from solving other similar scheduling problems.