The combination of traditional ant colony algorithm in solving the optimization process to consume a large amount of time, easily falling into local optimal solution and convergence is slow and other disadvantages, while also generating a lot of useless redundant iterative code, operation efficiency is low. Therefore, ant colony optimization algorithm is proposed. The algorithm based on genetic algorithm has the ability to search the global ant colony algorithm also has a parallel and positive feedback mechanisms. Changes in the use of genetic algorithm selection operator, crossover operator and mutation operator action to determine the distribution of pheromone on the path, the ant colony algorithm for feature selection using support vector machine classifiers for evaluating the performance characteristics of the feedback sub-Variorum And by changing the pheromone iteration, parameter selection and increase the local pheromone update feature nodes guided the re-combination. Simulation results show that the algorithm can effectively improve the accuracy, speed up the convergence, global optimization while enhancing the robustness and stability.%传统的组合优化蚁群算法在求解优化过程中要消耗大量的时间,极易陷入局部最优解和收敛速度过慢等弊端,同时还会产生大量无用的冗余迭代码,运算效率低；因此,提出一种遗传蚁群优化算法；该算法具备了遗传算法快速搜索全局能力的同时也具备了蚁群算法并行性和正反馈机制；利用遗传算法改变选择算子、交叉算子和变异算子操作来确定路径上信息素的分布,将蚁群算法用于特征选择,采用支持向量机分类器分类性能反馈用于评价特征子集解,并通过对改变信息素的迭代、参数选择和增加对信息素局部更新方式指导特征结点重新组合;仿真实验表明,该算法可以有效提高计算精度,加快收敛速度,优化全局最优解的同时增强了系统的鲁棒性和稳定性.
展开▼
