Research And Application On Flow Shop Scheduling Problem Based On Differential Evolution Algorithm

Job shop scheduling is the key technology of manufacturing informatization,and is important force to improve the core competitiveness of enterprises.How to use the advanced algorithms to solve practical industrial problems,rationally arrange production processes,and improve the production efficiency of enterprises has become a crucial issue.Differential evolution algorithm(DE)in Heuristic algorithm is a classical and effective algorithm to solve Job shop scheduling problem.In addition,it has good performance for various complex optimization problems.The improvements and applications of flow shop scheduling based on differential evolution algorithms are listed below.(1)An improved adaptive differential evolution algorithm(FMDE)is proposed.The mutation factor of the traditional differential evolution algorithm is a fixed constant,which leads to local optimum and premature population.Therefore,the mutation factor is changed into a decreasing function,called adaptive differential evolution algorithm.Although the improvement of the mutation factor ameliorates the performance of the algorithm,it doesn't take into account whether the population evolution is along the direction of the optimal solution.In order to solve the job shop scheduling problems effectively,a hybrid optimized algorithm based on modified adaptive differential evolution algorithm(FMDE)is proposed.FMDE is used to solve the mathematical model,and its performance is tested and analyzed by using the data of instance.The results show that FMDE can determine mutation rate adaptively,which enhances the probability of obtaining the global optimum.Compared with traditional differential evolution algorithms,FMDE converges faster and has higher performances.(2)A hybrid algorithm of particle swarm optimization and improved adaptive differential evolution(PSO_FMDE)is raised.Both particle swarm optimization(PSO)and differential evolution(DE)are based on population evolution.In order to avoid the population falling into the local optimal solution,we put forward a random mutation mechanism that avoids the stagnation of the population by generating random individuals.The performances of this algorithm are tested and analyzed by the test function.Compared with single algorithm,PSO_FMDE algorithm is easier to approach the global optimal solution with better performances.(3)The proposed FMDE algorithm and PSO_FMDE algorithm are applied to flow shop scheduling optimization.Flow shop scheduling is a typical production scheduling mode in process industry.Minimizing the makespan is set as the optimization target to solve the flow shop scheduling problem by transforming this issue into a mathematical model in which the FMDE and PSO_FMDE are used.At the same time,the performances of the flow shop scheduling is analyzed based on multiple standard test data by exploiting pascal language programming.The results show that the algorithms proposed in this thesis is easy to find the optimal scheduling scheme with excellent properties.In this thesis,two improved DE algorithms are proposed,and the effectiveness of algorithms is verified by the function optimization problem.The improved algorithms are applied to production scheduling optimization problems,and the results show that the algorithms have good performance in solving FSP problems.The research results in this thesis are helpful to promote the development of evolutionary algorithm and production scheduling theory,and have important theoretical and practical significance.