多目标优化的多种群混合行为二元蚁群算法

Multi-population binary ant colony algorithm with concrete behaviors for multi-objective optimization problem

针对二元蚁群算法在求解多目标问题时难以同时得到多个解和难以得到Pareto曲面的缺陷,使用多种群策略,改善算法的全局搜索能力,引入环境评价/奖励因子和蚁群混合行为搜索机制,提出了多种群混合行为二元蚁群算法。通过对几个不同带约束多目标函数的测试,实验结果表明该算法在保证全局搜索能力的基础上,拥有很好的多目标求解能力。

Aiming at solving the drawbacks of the original binary ant colony algorithm on multi-objective optimization problems：easy to fall into the local optimization and difficult to get the Pareto optimal solutions,Multi-Population Binary Ant colony algorithm with Concrete Behaviors（MPBACB） is proposed.This algorithm introduces multi-population method to ensure the global optimization ability, and uses environmental evaluation/reward model to improve the searching efficiency.Furthermore, concrete ant behaviors are defined to stabilize the performance of the algorithm.Experimental results on several constrained multi-objective functions prove that the algorithm ensures the good global search ability,and has better effect on the multi-objective problems.