Improved Slime Mould Algorithm Using Logical Chaos Perturbation and Reference Point Non-Dominated Sorting for Multi-Objective Optimization
Xuebing Cai, Zengyu He, Fei Cheng · IEEE Access · 2023
Slime Mould Algorithm (SMA) has been widely noticed by researchers for its powerful multi-point search capability and simple and feasible structure, and many advanced versions of SMA have also been proposed. However, most existing methods focus on the single-objective research domain, and research on multi-objective SMA remains relatively scarce; moreover, the basic SMA lacks powerful global search capability, and its extension to the multi-objective domain easily leads to the loss of solution diversity. Therefore, in this paper, we propose a general multi-objective SMA framework which utilises a logical chaotic single-dimensional perturbation mechanism to increase the search traversal of individuals in the decision space; secondly, a non-dominated sorting mechanism based on reference point is used to select more diverse solutions to participate in the next generation of evolution. Through experiments with seven advanced multi-objective algorithms such as CMOPSO, NSGA-II, NSGA-III, MOEAD, PSEA-II, SPEA-II and NSLS on 28 basis functions, the results show that multi-objective SMA achieves the best convergence, accuracy and diversity.