MOPOA: A New Multi-Objective Pufferfish Optimization Algorithm

Abstract Multi-objective optimization problems (MOPs) pose significant challenges due to the presence of multiple conflicting objectives. This paper introduces MOPOA, a novel Multi-Objective Pufferfish Optimization Algorithm inspired by the defensive behaviors of pufferfish in nature. MOPOA extends the original single-objective POA by incorporating Pareto dominance, an external archive for preserving non-dominated solutions, and a crowding distance mechanism to maintain solution diversity. The algorithm balances exploration and exploitation through biologically inspired phases simulating predator-prey interactions. To evaluate MOPOA’s performance, it was benchmarked against several state-of-the-art algorithms, including NSGA-III, MOPSO, MODA, and MOFDO, on two well-known test suites: the ZDT and CEC-2019 multi-objective functions. Results indicate that MOPOA not only achieves superior convergence to the Pareto front but also maintains high diversity and robustness across diverse optimization scenarios. These findings position MOPOA as a powerful and adaptive tool for solving complex real-world multi-objective problems.