A Two-Layer Genetic Algorithm Based on Matrix Coding for FJSP

To solve the problem of large number of redundant search existing in traditional genetic algorithm for solving Flexible Job Shop Scheduling Problem (FJSP), a two-layer genetic algorithm based on matrix coding for FJSP was proposed. The algorithm used double matrix structure to encode which included machine selection matrix and operation sorting matrix. Outer genetic algorithm is used to search for the best option of machine selection, which transform FJSP to JSP. Inner genetic algorithm is used to solve the sub-problem of JSP based on machine selection matrix. The effectiveness of proposed algorithm was verified by analyzing the results of benchmarks.