Scheduling For Flexible Flow-shop Problem Based On An Improved Genetic Algorithm

Flexible Flowshop Scheduling Problem is a kind of a simplified model, which abstracts from the field of real production scheduling. It is an extension of parallel machine and scheduling problem. The major characteristic is that there are parallel machines in some or all of the processes, which is very common in the process of flowshop manufacturing and it occupies a central position in the enterprise production management. Only by using outstanding scheduling according to the demand to allocate the manufacturing resources reasonably can the enterprises effectively improve the production efficiency, so as to improve their own competitiveness. But the inefficiency production scheduling made by workers is still existing among the most of enterprises. Therefore, the researching of intelligent algorithms on this problem has the extremely important theory significance and practical value.Genetic Algorithm (Genetic Algorithm, GA) is a kind of learn from biological natural selection and genetic mechanism of intelligent search algorithm, which solve the problem of simplicity and robustness of the characteristics and widely used in all areas of manufacturing. Still, in the process of algorithm development, it has some deficiencies. Such as the performance of the search efficiency is low, prematurely convergence, into the local extremum. According to the above reasons, this thesis has improved the genetic algorithm, and applied it to the flexible flowshop scheduling problem solving.In view of the method and the development about workshop scheduling, this thesis studies detailed about flexible flow-shop scheduling problems. According to the condition of workshop scheduling this thesis present flexible flow-shop scheduling problems with setup times and builds its mathematics modeling. Then using the triple matrix encoding method and taking the minimizing completion time as the objective function, this thesis designs an improved adaptive genetic algorithm with elite strategy improving traditional genetic algorithm which has slow convergence rate and was easily trapping in local optimum. The algorithm improves the crossover probability and mutation probability to make it changing along with the change of the fitness function, the algorithm which is proposed in this thesis made a simulation experiment on an example of a real flexible workshop comparing with traditional genetic algorithm. The result shows that the proposed algorithm was more effective than traditional algorithm in Scheduling results, algorithm convergence.