A hybrid optimization algorithm for multi-objective flexible job-shop

A hybrid optimization algorithm based on double-swarm particle swarm optimization (DPSO) algorithm and heuristic assignation(HA) algorithm is proposed to solve multi-objective flexible job-shop scheduling(FJSP) problem in this paper. In DPSO algorithm, the population is divided into two sub-populations and they are evolved with the different learning strategy respectively. The information is exchanged between the two parts and sizes of two sub-populations are changed according to corresponding iteration when population is again divided randomly into two parts in the appointed iteration. The reproduction strategy based on density of immune algorithm is introduced into DPSO algorithm to maintain the multiplicity of particle. DPSO algorithm can effectively avoid the premature convergence problem and improves the ability of searching a global optimum solution. DPSO-HA algorithm is applied to a set of benchmark problem instances. The simulation results have shown the effectiveness and feasibility of DPSO-HA algorithm for the multi-objective flexible job-shop scheduling.