A Novel Decomposition-Based Multi-Objective Evolutionary Algorithm with Dual-Population and Adaptive Weight Strategy

Multi-objective evolutionary algorithms mainly include the methods based on Pareto dominance relationship and decomposition. The method will produce a large number of non-dominated individuals with increase in population size or objectives, resulting degradation algorithm performance. Although decomposition is not limited by it does perform well complex front due to fixed setting weight vector. In this paper, we combined these two different approaches proposed Multi-Objective Evolutionary Algorithm Decomposition Dual-Population Adaptive Weight strategy (MOEA/D-DPAW). vector adaptive adjustment used periodically change evolution process, information interaction between populations enhance neighborhood exploration mechanism improve local search ability algorithm. experimental results 22 standard test problems such as ZDT, UF, DTLZ show that paper has better performance than mainstream multi-objective recent years, solving two-objective three-objective optimization problems.