Solving Flexible Job Scheduling Problem Based On Genetic Algorithm

Manufacturing industry is closely related to people's life,and its development can directly influence the national comprehensive strength.Not only front-end manufacturing technologies but also management technologies are required for manufacturing enterprises to maintain their core competitiveness at all times.Enterprise scheduling remains the core as well as the trouble of enterprise management.The formulation of production scheduling is the most important task of manufacturing operation management.The results can strongly affect the profit of the enterprise,the utilization efficiency of resources,whether the product can be delivered on time or not,and so on.The actual flexible job shop scheduling environment allows all processes to arbitrarily choose one machine in the entire set of machines.Flexible job scheduling problem(FJSP)can allocate resources flexibly according to the resource load,improve the flexibility of processing,and get closer to the actual production environment.However,it increases the search scope of feasible solution and the complexity of the problem,in addition,the history of the research on FJSP is short,thus there are still some outstanding difficulties to be solved in the model and the solution strategy for flexible job scheduling problem.As a swarm intelligence algorithm,genetic algorithm,which is easy to operate,has better search ability and robustness in solving FJSP,but there are still some shortcomings,such as slow convergence speed,easy early maturity and insufficient population diversity.In view of the shortcomings of genetic algorithm in solving FJSP,the following improvements are made in this paper.Considering the diversity of population and process sequence constraints,a global random selection mechanism(GRS mechanism)for initializing chromosome segments is proposed.When the algorithm is in the local optimal state,a re-activation mechanism is added to update the population and adjust its population diversity to the initialization state.Then a improved genetic algorithm is proposed.The population independence ratio is designed to measure the diversity of the population,and the niche technology is introduced into GA.A preselection mechanism based on niche technology is added to GA after each genetic updating operation.and a improved the niche genetic algorithm is proposed.In this paper,the two improved algorithms are applied to 11 sets of standard test cases,which verify the reliability and effectiveness by compared with other algorithms.Then,the improved algorithm is applied to three practical flexible job scheduling systems,which proves that the improved algorithm is effective in practical examination.