双概率原对偶遗传算法与蚁群算法的融合研究

Combination of Double-Probability Primal-Dual Genetic Algorithm and Ant Algorithm.Computer Engineering and Applications

双概率原对偶遗传算法(DPPDGA)是原对偶遗传算法的一种改进型算法,赋予各基因位值不同的对偶映射概率,增强算法种群的多样性,具有更好的全局寻优性能,但不能利用系统中的反馈信息,出现无为的冗余迭代。最大最小蚁群算法(MMAS)则能够很好的利用系统中的反馈信息,通过对信息的累积更新寻求最优解,但初始信息素的匮乏制约着MMAS的求解效率。本文将两种算法进行融合,克服自身缺陷,优势互补。通过MATLAB仿真测试可知,该融合算法表现出求解精度高、稳定性强、全局搜索性能优的特点。

A new improvement of primal-dual Genetic Algorithm that its gene has a variety of probability,called Double-Probability Primal-Dual Genetic algorithm（DPPDGA）, is presented. It improves the diversity of population and the ability of global searching quickly, but it cannot make use of system feedback information that it iterates redundantly. Max-Min Ant algorithm（MMAS）can make full use of system feedback information and seeks the best result by updating the pheromone. But the efficiency is restricted due to the lack of initial pheromone. A new combination of algorithms of DPPDGA and MMAS has been put forward, which utilizes the characteristics of the two algorithms. The MATLAB simulations show the new algorithm is more efficient, more stable and more precise.