Abstract
Genetic programming, as a hyper-heuristic approach, has been successfully used to evolve scheduling heuristics for job shop scheduling. However, the environments of job shops vary in configurations, and the scheduling heuristic for each job shop is normally trained independently, which leads to low efficiency for solving multiple job shop scheduling problems. This paper introduces the idea of multitasking to genetic programming to improve the efficiency of solving multiple dynamic flexible job shop scheduling problems with scheduling heuristics. It is realised by the proposed evolutionary framework and knowledge transfer mechanism for genetic programming to train scheduling heuristics for different tasks simultaneously. The results show that the proposed algorithm can dramatically reduce the training time for solving multiple dynamic flexible job shop tasks.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have