Abstract:
A cooperative co-evolutionary genetic algorithm is proposed for a distributed flexible job shop scheduling problem with the optimization objective of minimizing the maximum completion time. A decoupled encoding of factory assignment and operation sequencing is used, based on machine load decoding as well as initializing the population based on factory load, so that the algorithm iteratively searches in a better solution space. Using the divide-and-conquer idea, the problem is decomposed into multiple sub-problems, and a random collaboration mechanism is used to promote the subpopulations to co-evolve and improve the global exploration capability. Multiple local perturbation strategy based on key factory is used to improve the local exploitation capability. Experiments are conducted on benchmark instances and compared with other algorithms to verify the effectiveness of the proposed algorithm.