GOM-MMOEA: Multimodal multi-objective evolutionary algorithm based on global orchestration mechanism

Abstract

The core challenge of multimodal multi-objective optimization lies in identifying and discovering multiple equivalent sets of Pareto-optimal solutions, thereby offering diverse options for decision-makers. However, most existing algorithms suffer from premature convergence when tackling such problems. This issue often arises due to inadequate population diversity and ineffective global exploration mechanisms during the search process, which causes the algorithm to become trapped in local optima and hinders the exploration of other promising regions in the decision space. To address this challenge, this paper proposes a multimodal multi-objective evolutionary algorithm based on a global orchestration mechanism. First, the algorithm constructs and dynamically updates an orchestration vector to guide the search toward optimal solutions and accelerate population convergence. Second, an orchestration vector update strategy is designed to gradually diminish the influence of inferior solutions, thereby preventing convergence to local optima. During the early stages of evolution, larger increments are applied to high-quality solutions to speed up convergence, while these increments are gradually reduced over time to promote global exploration. Finally, a novel parent selection mechanism is introduced, which dynamically adjusts selection probabilities to optimize the search process while preserving population diversity. Moreover, the algorithm adopts a triple population synergistic orchestration method that simultaneously considers both the objective and decision spaces. Experimental results demonstrate that the proposed algorithm outperforms several state-of-the-art methods across a range of benchmark test problems.