摘要：In this paper we propose a new Copula-based Estimation of Distribution Algorithm, to solve Many-objective optimization problem and to get new optimal solutions in very court time. Our algorithm uses the proprieties of Copula and exploits their statistical properties to make new solutions using the founded optimal solutions through the estimation of their distribution. The first step of the proposed Copula-based Estimation of Distribution Algorithm (CEDA-SVM) is taking initial solutions offered by any MOEA (Multi Objective Evolutionary Algorithm), and then creates Copulas to estimate their distribution, and we use Support Vector Machine (SVM) to learn the Pareto solutions model; those Copulas will be used to generate new solutions and SVM to avoid the expensive function evaluations. The idea of using the estimated distribution of the optimal solutions helps CEDA-SVM to avoid running the optimizer ( MOEA ) every time we need new alternatives solutions when the founded ones are not satisfactory. We tested CEDA-SVM on a set of many-objective benchmark problems traditionally used by the community, namely DTLZ (1, 2, 3, and 4) with different dimensions (3, 5, 8, 10, and 15). We used CEDA along with MOEA/D-Schy and MOEA/D-BI as two examples of MOEA thus resulting in two variants CEDA-MOAE/D-Scy and CEDA-MOEA/D-BI and compare them with MOEA/D-Schy and MOEA/D-BI. The results of our experiments show that, with both variants of CEDA-SVM, new solutions can be obtained in a very small time compared to the other algorithms.
关键词：Multiobjective optimizationmany-objective optimizationestimation of distribution algorithmSVMCopula
