Research On Multi-objective Differential Evolution And It's Applications

Multi-objective optimization algorithm is the hot issue in the field of study.In naturalscience and social science,most optimization problems belong to the optimization of complex multi-objective,according to this kind of problem,the traditional optimization method is hard to get the better result,so the optimization algorithm methods are used to solve the problems effectively.The basic principle of using evolutionary algorithm to solve complicated multi-objective optimization problem is that can parallel search multiple targets for the evolution of the population.So we can find the optimal solution of multi-objective problem,and also reduce the probability of getting into the local optimum.A large number of experiments show that the evolutionary algorithm has been obtained the widespread attention and the application in the research field.Differential evolution algorithm is a kind of new intelligent optimization algorithm,its principle simple,convenient operation,less controlled parameter and strong robustness.Therefore,DE algorithm already obtained attention of the international scholars.Although DE algorithm has obvious advantages compared with other evolution algorithms,scholars found that the differential evolution algorithm also has many places need to improve.So it is necessary to research on the differential evolution algorithm,and make it can solve more complex optimization problems in practice.The main contribution of this paper can be summarized as follows:(1)The parameters of DE algorithm are adjusted adaptively,which combined with swarm optimization algorithm that introducing the Tent chaos and inertia weight adaptive adjustment operation.The results show that the proposed new hybrid algorithm is not easy to appear "premature" phenomenon,and has better search accuracy.(2)Using niche technology and rapid replacement algorithm of clustering crowd out for fitness and environmental selection operation,the adaptive mutation operator is set to the parameters.Thus,a adaptive multi-objective differential evolution algorithm based on the strength of the Pareto was bulit.The result show that the adaptive strength Pareto based multi-objective differential evolution algorithm has good distribution and convergence.(3)The parameter adaptive adjustment operation and the feasibility rules were introduced in the basic multi-objective differential evolution algorithm,and the improved algorithm is applied to multi-objective optimization problem of portfolio.In the simulation experiments,using the real data analysis of the stock market,the improved multi-objective differential evolution algorithm can effectively guide the investment decision,and perform better in termsof real yields.