SMSP-EMOA: Augmenting SMS-EMOA with the prospect indicator for multiobjective optimization

This paper studies a new evolutionary multiob-jective optimization algorithm (EMOA) that leverages quality indicators in parent selection and environmental selection operators. The proposed indicator-based EMOA, called SMSP-EMOA, is designed as an extension to SMS-EMOA, which is one of the most successfully and widely used indicator-based EMOAs. SMSP-EMOA uses the prospect indicator in its parent selection and the hypervolume indicator in its environmental selection. The prospect indicator measures the potential (or prospect) of each individual to reproduce offspring that dominate itself and spread out in the objective space. It allows the parent selection operator to (1) maintain sufficient selection pressure, even in high dimensional MOPs, thereby improving convergence velocity toward the Pareto-optimal front, and (2) diversify individuals, even in high dimensional MOPs, thereby spreading out individuals in the objective space. Experimental results show that SMSP-EMOA's parent selection operator complement its environmental selection operator. SMSP-EMOA outperforms SMS-EMOA and well-known traditional EMOAs in optimality and convergence velocity without sacrificing the diversity of individuals.