A multiobjective evolutionary algorithm based on surrogate individual selection mechanism

Abstract

Recently, classification-based preselection (CPS) strategy for evolutionary multiobjective optimization has been found to be very effective and efficient for solving complicated multiobjective optimization problems (MOPs). However, this strategy can only classify the candidate solutions into different categories, but it is difficult to find out which one is the best. In order to overcome this shortcoming, we propose a surrogate individual selection mechanism for multiobjective evolutionary algorithm based on decomposition. In this mechanism, we get the best one from candidate solution set by surrogate model, which mitigates the risk of using CPS strategy. Furthermore, we generate candidate solution set through a new offspring generation strategy, which can improve the quality of the candidate solutions. Based on typical multiobjective evolutionary algorithm MOEA/D, we design a new algorithm framework, called MOEA/D-SISM, through integrating the proposed surrogate individual selection mechanism. We compare MOEA/D-SISM with other state-of-theart multiobjective evolutionary algorithms (MOEAs), and experimental results show that our proposed algorithm obtains the best performance.

[Retracted] Multimodal Emotion Recognition Model Based on a Deep Neural Network with Multiobjective Optimization

“…A multiobjective optimization problem (MOP) is a kind of challenging and complex optimization problem. Because the optimization goals conflict with each other, it is extremely difficult to obtain a single global optimal solution, so it is a set of compromise Pareto optimal solutions [24,25]. In recent decades, many similar optimization algorithms have appeared, such as PEAS [26], SPEA2 [27], NSGAII [28], MOEA [29], MOEA/D [30], IBEA [31], and HypE [32].…”