The Research On Multi-objective Evolutionary Optimization Algorithms Based On Diversification Strategy

With excellent performance in solving multiobjective optimization problems, evolutionary algorithm has become a hot research topic in the field of multi-objective optimization. For improve the speed of convergence, robustness and distribution of multi-objective evolutionary algorithm. This paper designed and improved a series of evolutionary operators based on basic theory of evolutionary algorithm. Meanwhile, the diversification strategy acts as a breakthrough point for all innovations proposed here. The main contents are as follows:Firstly, this paper introduces the basic concept of multi-objective optimization problem and its research significance. And then the current research status of multi-objective evolutionary algorithms was reviewed. Meanwhile, some relevant basic theories were expounded, including the design target and design key points of the multi-objective evolutionary algorithm. The benchmark problems and performance metric are also presented in the end.Secondly, this paper proposed a novel multi-objective evolutionary algorithm based on the scatter-interference-mechanism. The new approach overcame the location independence of the crowding-distance based method, which is known as a weakness. In order to guarantee quality of the solution set, meanwhile, enhance the population diversity and compensate for the restrictions on search space, the strategy of interference set is also proposed.Thirdly, by analyzing the importance of evolutionary direction for evolutionary algorithms, the approximate direction and distribution of population direction were made full use to guide and control the evolution of the population. By making full use of the characteristics of non-dominated and dominated sets, a new distributed-guiding strategy was propose to better guide the evolution of individuals.Fourthly, a hybrid scatter-search based multi-objective evolutionary algorithm was proposed. The scattered search mechanism was incorporated for the reason that it can strike a good balance between the quality of solution sets and its distribution, both of which are primary issue to be considered.Finally, the main research contents are summarized at the end of the thesis with an expectation for further study and research.