Decomposition Based MOEA with Unique Subregions and Stable Matching

The Multiobjective Evolutionary Optimization algorithms (MOEAs) have attracted lots of attention and have been used for resolving engineering problems, such as production scheduling, logistics planning, and intelligent storage. This paper focuses on the modification of the Stable Matching-based Selection (STM) in Evolutionary Multiobjective Optimization (MOEA/D-STM). Generally speaking, the mechanism would strike an equilibrium between the preferences among scalar problems and solutions. Nevertheless, the initial stable matching model may have a high weakness of mismatching with a certain solution. To address the drawback, this paper proposes the Decomposition Based MOEA with Unique Subregions and Stable Matching (MOEA/D-RM). MOEA/D-RM decomposes the multiobjective optimization problem into a number of various subregions for remedying the loss of diversity and introduces the incomplete preference lists into the STM model for repairing the loss of convergence. The effectiveness and competitiveness of MOEA/D-RM are corroborated and compared with several evolutionary multiobjective optimization algorithms.