A new human-based offensive defensive optimization algorithm for solving optimization problems

Abstract

A novel human-inspired metaheuristic algorithm, termed Offensive Defensive Optimization, has been introduced to address single-objective optimization problems. This algorithm draws inspiration from the varied strategies utilized by players in board games, emulating and conceptualizing offensive and defensive behaviors within a hybrid search framework. The integration of mixed search behaviors facilitates a more efficient exploration and exploitation of the search space, thereby enhancing the algorithm’s capability to surmount local minima. The algorithm was evaluated using the benchmark test suites from the Congress on Evolutionary Computation (CEC) 2017 and 2022, in addition to two real-world engineering design problems. In comparison to eight well-established metaheuristic algorithms, the proposed method demonstrated superior performance in 80% of the CEC2017 cases and 72% of the CEC2022 cases, with statistically significant improvements. The results further indicate that the proposed algorithm exhibits satisfactory convergence efficiency, along with robust exploration and exploitation capabilities, while maintaining a balanced equilibrium between these two processes. Additionally, the outcomes of the engineering design problems suggest that the proposed algorithm effectively manages optimization tasks, demonstrating clear superiority and enhanced competitiveness.