改进的遗传蚁群混合算法求解多维0/1背包问题

Modified Genetic Ant Colony Hybrid Algorithm for Solving Multidimensional 0/1 Knapsack Problem

针对传统遗传蚁群混合算法求解精度低、收敛速度慢等缺陷,设计了一种改进的遗传蚁群混合算法,该算法选择部分优秀蚂蚁进行遗传算法寻优并更新全局信息素,其它蚂蚁采用蚁群算法寻优,并更新局部信息素。其中对传统遗传算法的交叉和变异操作进行了改进,并在蚁群算法的运行过程中引入概率和为u的轮盘赌方式以减少计算量、采用禁忌表交换策略以及信息素的混沌更新策略来增强种群多样性,避免陷入局部最优。实验结果表明,该算法在求解精度和收敛速度方面都有明显提高。

To overcome the problems of low precision and convergence speed of traditional genetic and ant colony hybrid algorithm,an modified genetic and ant colony hybrid algorithm is designed. The modified algorithm selects some excellent ants for genetic algorithm optimization and update the global pheromone. Other ants use ant colony algorithm to optimize and update the local pheromone. Meanwhile,some improvements at crossover operation,mutation operation are proposed. In the process of ant colony algorithm,the roulette way that the sum of probabilities is u are introduced to reduce the computational cost. The tabu exchange strategy and the pheromone update strategy of the chaotic are used to enhance the diversity of the population and avoid the local optimum. The experimental results show that the algorithm has improved the precision and the convergence rate.