An Improved Genetic Algorithm for Flexible Job-Shop Scheduling Problems

Abstract

The flexible job-shop scheduling algorithm with the high computational complexity is important in both fields of production management and combinatorial optimization. We developed an improved genetic approach for the flexible job-shop scheduling problem (FJSP), since it is quite difficult to achieve an optimal solution to this problem within reasonable computation time. A chromosome representation combines both routing and sequencing information is represented with genetic algorithms (GAs) commonly optimize the problem by minimizing an objective, makespan. Our algorithm selects the chromosome according to different objectives, and crossover each other to maintain the species diversity. New crossover and the mutation operators utilize the critical path with certain probability to modify assignment and sequence for preventing being trapped in a local optimum. The results obtained from the computational study have shown that the proposed algorithm is an effective approach for the FJSP.