The Research On High-objective Optimization Evolutionary Algorithm Based On Dynamically Adaptation Of Reference Vector

Classical multi-objective evolutionary algorithms(MOEAs)have been proven to be inefficient for solving multi-objective optimizations problems when the number of objectives increases due to the lack of sufficient selection pressure towards the Pareto front(PF).In addition,maintaining a good balance between convergence and diversity is crucial to the performance of MOEAs.This poses a great challenge to the design of MOEAs.To cope with this problem,researchers have developed reference-point based methods,where some well-distributed points are produced to assist in maintaining good diversity in the optimization process.However,the performance of MOEAs be severely affected by the Pareto front(PF)of the multi-objective optimization problems(MOPs)during the searching procedure.This paper proposes a reference-point-based adaptive method to study the PF of MOPs according to the candidate solutions of the population.This adaptive method is to adjust the relative position of reference points to deal with MOPs that have concave or convex PF,and use a parameter ? to control the hyperplane shape of the reference points according to the objective value of whole population so that the distribution of solutions can be improved.The new clustering method is designed to balance the niche-preservation operation in Multi-objective optimization according to the reference points.We divide a population into a number of niches by value of the proportion and angle,and then select a better solution in the niche.In this method,convergence and diversity are primarily measured by proportion and angles,which to strengthen the convergence to the PF as well as maintain a good diversity of the population during environmental selection.The six state-of-the-art MOEAs,HYPE?SPEA2+SDE?MOEA/D?MOEA/DD? NSGA-III and RVEA is selected as the comparison algorithms.The ZDT,DTLZ and WFG MOPs have various features,such as having a linear,multi-modal,concave,discontinuous,or degenerate PF.Compared with six MOEAs,the proposed algorithm shows highly competitive effectiveness on MOPs in terms of diversity and convergence.In addition,the parameter ? have the ability to adjust dynamically the hyperplane shape of the reference points according to the objective value of whole population with the evolutionary,and obtain a good balance between convergence and diversity.