Abstract:

We propose an improved fruit fly optimization algorithm named dynamic double population fruit fly optimization algorithm (DDFOA) to solve the flexible job shop scheduling problem (FJSP). The DDFOA adopts a self-adaptive moving step length and divides the population into two parts, in which the backward sub population focuses on global search, and the advanced sub population is responsible for local search. Then, we design an appropriate code conversion program for the FJSP. The convergence of the proposed algorithm is proved. Simulation results on several benchmarks show that the DDFOA is an effective approach for solving the FJSP.

Key words: flexible job shop scheduling, double population, fruit fly optimization algorithm, neighborhood search