Research Of Flexible Job Shop Scheduling Problem Based On Hybrid Discrete Particle Swarm Optimization Algorithm

Production scheduling is capable of improving the production efficiency, reducing the cost and energy consumption, and is benefit to keep the long-term developments of enterprises. Hence production scheduling has been a hot research topic of the computer integrated manufacturing field in the past few decades. As an important branch of production scheduling, flexible job shop scheduling problem is more close to the real production process but is much more difficult to solve. Thus the study of flexible job shop scheduling problem is of significant research interests both practically and theoretically.In this paper, several kinds of flexible job shop scheduling problem are studied by proposing the corresponding hybrid discrete particle swarm optimization algorithms. The algorithms are based on the standard discrete particle swarm optimization algorithm and are modified by improving the particle position updating strategy and presenting various local searching schemes suitable for the problems. The main research works of this paper are summarized as follows:(1) With a modified particle position updating strategy and a machine-load-based simulated annealing technique, a hybrid discrete particle swarm optimization algorithm is proposed to solve a single-resource-single-objective flexible job shop scheduling problem with machine being the constrained resource and production cycle time being the optimization objective. The hybrid algorithm effectively compensates for a disadvantage of particle swarm optimization algorithm in solving the flexible job shop scheduling problem, which is prone to result in infeasible solutions. Moreover, the algorithm not only guarantees the convergence of the particle to feasible solutions, but also improves the searching ability. Simulation examples show that the proposed algorithm is feasible and effective.(2) Based on Pareto dominance relationship, a hybrid discrete particle swarm optimization algorithm is proposed to solve a single-resource-multi-objective flexible job shop scheduling problem with a modified Baldwinian learning mechanism and simulated annealing technique being the local searching scheme. The optimization objectives are production cycle time, total machine load and the maximal single machine load, respectively. Besides the improvements of the local searching ability, the algorithm also uses mixed initialization way of particle swarm to enhance the starting point of searching and introduces an external archive to prevent the loss of non-dominated solutions. The simulation results show that the algorithm achieve better performance than some other methods in both convergence and diversity of the non-dominated solutions.(3) For a flexible job shop scheduling problem with both machine and worker being the dual constrained resources, and with production cycle time being objective, a hybrid discrete particle swarm optimization method is presented by combining modified discrete particle swarm optimization algorithm and a simulating annealing technique based on variable neighborhood. The algorithm is also modified in the initialization of the particle swarm, position updating mechanism of particles and variable neighborhood selection to avoid the occurrence of infeasible solution and premature convergence. Dual-resource examples are used to demonstrate the effectiveness and feasibility of the proposed algorithm.(4) A dual-resources-multi-objective flexible job shop scheduling problem is studied with machine and worker being the constrained resources and with production cycle time and production cost being the optimization objectives. Based on the Maximin fitness function, a hybrid discrete particle swarm optimization algorithm is proposed to solve the problem. The algorithm improves the ability of local searching effectively. Moreover, a simple mixing strategy is proposed to pruning the external archives, which guarantees the algorithm efficiency in the running time. Illustrative examples are simulated to show the effectiveness of the presented algorithm.Finally, a flexible job shop scheduling system with B/S structure is designed with the hybrid algorithms proposed to solve the flexible job shop scheduling problems in this paper. Two practical examples in a paper cone shop are used to illustrate the effectiveness of the proposed method. Then the research work of this paper are summarized and prospected.