最適化問題に対する三層ハイブリッド探索アルゴリズム

抄録

This paper proposes a triple-layer hybrid optimization algorithm that integrates Monte Carlo sampling, Genetic Algorithm, and Ant Colony Optimization. The proposed method enables adaptive cooperation among probabilistic exploration, evolutionary search, and pheromone-based reinforcement, thereby balancing exploration and exploitation dynamically. Benchmark experiments on Rastrigin, Rosenbrock, and Ackley functions demonstrated that the proposed algorithm achieved the best solution quality among conventional single algorithms, while successfully avoiding premature convergence. These results indicate that the proposed hybrid approach effectively enhances global search capability and robustness in continuous optimization problems.