Evolutionary algorithm for solving multi-objective hybrid flow-shop scheduling problem

To solve multi-objective hybrid flow-shop scheduling problems, a new evolutionary algorithm was presented which could dynamically adjust the fitness assignment in optimizing scheduling process. Matrix code was used to describe the scheduling solutions of multi-phased parallel machines, and the improvement degree for each Pareto solution was measured by combining optimization model. Then the fitness of each individual was determined with a selective weight approach. Thus the adaptive pressure to proper direction could be obtained. The performance of this algorithm with benchmark problems and practical optimization example were analyzed, and experimental results showed that the algorithm was more effective than the existing ones for multi-objective problems with large dimensionality, and it could converge to satisfactory solutions at a high speed.