MMFOSOM: A Novel Multi-objective Moth–Flame Optimization Algorithm Based on Self-Organizing Mapping

An optimization technique seeks to identify the best solution for a given problem. When the problem involves a single objective function, the optimal solution yields the best value for that objective. However, in cases with multiple objectives, selecting the best solution becomes more complex, as these objective functions often conflict with one another. For such multi-objective optimization problems, relying on analytical or exact methods is often not feasible. This paper proposes a modified MFO called multi-objective moth–flame optimization based on self-organizing mapping (MMFOSOM) to solve multi-objective optimization problems. For attaining high-quality initial moths, dual opposition-based learning mechanism is employed to generate two distinct populations, referred to as the exploration moth (EM) and the auxiliary moth (AM). In order to enhance the search capabilities of the algorithm, each population is assigned a SOM network, the adaptive neighborhood structure of which contributes to forming modal clusters in the mapping, organizing similar solutions together. In order to select the excellent individual as the training data for SOM, a novel environmental selection mechanism is introduced. To validate the effectiveness of the algorithm, experiments will be conducted on a set of multi-objective optimization test functions. The experimental results indicate that the proposed algorithm can effectively address multi-objective optimization problems.