Hybrid Differential Evolution And Application To Production Scheduling

The scheduling problem is the center of manufacturing system operation management,and is widely used in industrial production,automation,and vehicle and ship scheduling.The shop scheduling problem is based on the constraints of the resources(time labor costs,raw materials,etc.),and the decision-maker allocates the processing operations to related procedures quantitatively,and configures the processing seq uence of the operations to achieve the optimal arrangement under the constraints.An excellent scheduling strategy can achieve better resource allocation,improve the efficiency of the production system,and achieve economic benefits.Therefore,research on scheduling issues has become a hot topic.This paper presents a multi-objective hybrid differential optimization algorithm.The hybrid algorithm absorbs the advantages of differential evolution vector operations and dynamically adjusts the search direction based on historical records.In order to overcome the shortcomings,which the differential evolution algorithm can easily fall into a local optimum due to the small population diversity.Through a mixed sampling strategy,the distribution information of the solution set is obtained to design the differential evolution mutation operator,so that the hybrid algorithm can improve the convergence or distribution in a favorable direction,so as to make up for the defect that is trapped in the local optimization and the algorithm can quickly Pareto forward.In this paper,firstly,the multi-objective hybrid differential optimization algorithm is verified on the classical benchmark problem sets,and compared with the traditional excellent algorithms.According to the disadvantages,different alternative individual selection methods are proposed in the differential evolution process,and comparative data analysis is performed.Finally,the mathematical model is built for the double-objective flow shop scheduling problem with the maximum completion time and the total flow time.The improved mutation operator is used to increase the ability of the algorithm's global search.The Benchmark Problems test results show that the hybrid differential evolution algorithm has obvious advantages in convergence performance and distribution performance for other multi-objective evolutionary algorithms.In the flow shop scheduling problem simulation experiment,the hybrid algorithm is more suitable for the flow shop scheduling problem.By comparing the different mutation operators,the alternative individual selection method based on global considerations is more competitive on the final elite solution set.The multi-objective hybrid differential evolution algorithm is superior to th e traditional evolutionary algorithm in testing the flow shop scheduling problem and Benchmark Problems.It also achieves good results after improving the mutation operator.