摘要
TSP是组合优化问题中著名的NP-hard问题。针对粒子群算法求解离散的TSP问题收敛速度慢,求解精度低,易于陷入局部最优和模拟退火算法的性能与参数初始值有关及参数敏感等不足,提出了将改进的粒子群算法作为全局搜索策略,改进的模拟退火算法作为局部搜索策略的文化基因算法。介绍了两种算法的协同方法,定义了局部搜索邻域的确定以及在新种群产生中引入自组织随机移民策略。仿真结果表明,改进算法在求解TSP问题中具有很快的收敛速度,且能搜索到最优解。
The TSP problem is famous NP - hard problem in combinatorial optimization. The this paper, we put forward an improved particle swarm optimization （pso） as a global searching and a simulated annealing algorithm as a local searching, which we called Memetic algorithm. We introduced two algorithms＇ synergism, defined the deter- mination of neighborhood and introduced self - organizing random immigrants strategy to generate new population. The simulation results show that the proposed algorithm has fast convergence speed in solving TSP problem, and can search the optimal solution.
出处
《计算机仿真》
CSCD
北大核心
2015年第2期284-287,358,共5页
Computer Simulation
基金
自治区自然基金(2012211A003)
