求解CVRP的改进混合蛙跳算法研究

摘要

为了求解带有容量约束的车辆路径问题(CVRP),在建立CVRP数学模型的基础上,提出了改进的混合蛙跳算法(SFLA),并设计了新的初始群体构造方法.改进后的SFLA采用实数编码方式,融入自适应差分扰动机制及混沌局部搜索策略到局部搜索过程中,在保持SFLA全局收敛性的同时,增强了算法跳出局部最优解的能力,加快了算法收敛速度.通过与其他三种算法进行对比实验,结果表明了改进的SFLA在求解CVRP上的有效性和顽健性.%To solve CVRP,this paper proposed a modified SFLA, which based on the mathematical model of CVRP, and de-signed a new method for constructing initial population. The modified shuffled frog leaping algorithm adopted real-coded pat-terns. Then it combined with adaptive differential disturbance and chaotic local search strategy in local searching. It enhanced the ability to escape from local optima and sped up the convergence of SFLA, meanwhile, maintained global convergence of SFLA. Experimental results indicate the effectiveness and robustness of the modified shuffled frog leaping algorithm in solving CVRP via comparing with three other algorithms.