A decision variable classification-based cooperative coevolutionary algorithm for dynamic multiobjective optimization

This paper proposes a new decision variable classification-based cooperative coevolutionary algorithm, which uses the information of classification to guide search process, for handling dynamic multiobjective problems. In particular, variables are divided into two groups: convergence (CS) and diversity (DS), different strategies introduced optimize these groups. Two kinds subpopulations used in proposed i.e., that represent DS CS. evolution coevolution CS is carried out through genetic operators, gradually merged DS, optimized global space, based on an indicator avoid becoming trapped local optimum. Once change detected, prediction method introduction approach adopted get promising population with good environment. The algorithm tested 16 benchmark optimization problems, comparison state-of-the-art algorithms. Experimental results show very competitive optimization.