Font Size: a A A

Research On Two Kinds Of Improved Evolutionary Computing And Swarm Intelligence Multi-objective Optimization Algorithms

Posted on:2020-08-12Degree:MasterType:Thesis
Country:ChinaCandidate:S W XuFull Text:PDF
GTID:2438330626963976Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
With the development of the society,optimization problems have received more and more attention.Multi-objective optimization problem is a common optimization problem in daily life,and it is of great significance in the field of scientific research and engineering application.Evolutionary algorithm and swarm intelligence optimization algorithm are two important intelligent optimization algorithms to solve multi-objective optimization problems.In terms of weight adjustment mechanism and population renewal strategy,this paper improves the multi-objective evolutionary algorithm based on decomposition and artificial bee colony algorithm,which are the represents of two intelligence optimization algorithms.Moreover,simulation experiments of new algorithms on corresponding multi-objective optimization problemss were carried out with Matlab software.The main contents of this paper can be summarized as the following two points:(1)A modified multi-objective evolutionary algorithm based on decomposition with an adaptive weight adjustment strategy is proposed for multi-objective optimization problem of Pareto Front with complex geometry(MOEA/D-BPO).MOEA/D-BPO mainly includes elite individual mechanism,archives mechanism and adaptive weight vector adjustment strategy.Firstly,using Tchebycheff approach to decompose the multi-objective problem,and then updating population and producing new offspring group,updating elite groups and archives,finally when the condition of the adaptive weighting vector is satisfied,updating and adjusting weight vector.Through the comparison experiment,it is concluded that the MOEA/D-BPO has a good performance in solving the type target problem.(2)For complex numerical problems,a surrogate-assisted multi-swarm artificial bee colony is proposed(SAMSABC).SAMSABC mainly includes K-means clustering method,orthogonal method,modified update process and fitness estimation strategy.Firstly,the whole population was reclassified,and then the optimal individual in each subpopulation and the individual closest to the optimal individual were found.The possible potential optimal position was calculated by using orthogonal design method and fitness function,and finally the individuals were updated by using the improved individual updating mechanism and possible potential optimal position.Through the comparison experiment,it is concluded that theSAMSABC has a good performance in solving the type target problem.
Keywords/Search Tags:intelligent optimization algorithm, multi-objective optimization problem, decomposition, weight adjustment, artificial bee colony algorithm
PDF Full Text Request
Related items