For extending the ability of multi-objective optimization for continuous functions for the ant colony algorithm, this paper proposed an improved ant colony algorithm (MO-FHACO) based on the foraging-homing mechanism inspired by the natural ant colonies who laid the different pheromones. The pheromones were divided into two kinds, i. e. the nest pheromone and the food pheromone, on the path from the nest to the food resource. Therefore, it built the foraging-homing mechanism to find the function optimal value. According to thi function test of BNH and TNK, results show that MO-FHACO has the best multi-objective function optimization ability comparison with other intelligence algorithms, if Pareto frontier is continuous. And if Pareto frontier is discontinuous, MO-FHACO still can get good Pareto optimum values. So MO-FHACO is an efficient multi-objective function optimization algorithm.%受自然界蚂蚁的觅食—返巢生物学特征启发,同时深入了解蚂蚁信息素成分,提出了一种能够解决函数多目标优化问题的改进蚁群算法——多目标觅食—返巢机制连续域蚁群算法(MO-FHACO).该算法与传统蚁群算法相比,将信息素分为蚁巢信息素和食物信息素,并根据不同信息素设立了不同的释放和寻优机制.通过BNH和TNK问题验证,MO-FHACO算法在Pareto最优前端连续的情况下具有极佳的多目标优化能力;在Pareto最优前端不连续的情况下,也能得到较多且散布性较好的Pareto最优解.因此,MO-FHACO算法是一种有效的函数多目标优化算法.
展开▼
