基于改进模拟退火算法求解TSP问题

摘要

By analyzing the principles and the shortcomings of traditional simulated annealing algorithm, the improved simulated annealing algorithm is proposed for the characteristic of TSP. The traditional simulated annealing algorithm producing the new solution is random, when the parameter is not accurate. It cant lead to the optimal solution; the traditional simulated annealing algorithm is easy to lose the optimal solution of current. Put forward a new method to select the initial solution and generation the solution, the method of the improved the current solution and adding memory function for optimal solution of current. Experimental test results show that the new algorithm has faster convergence and stability than traditional simulated annealing algorithm.%对传统模拟退火算法的原理和不足进行分析,针对TSP问题的特点提出了改进的模拟退火算法.就传统模拟退火算法生成新解的随机性太强、参数设置不当不能搜索到全局最优解、容易丢失当前最优解等问题提出了新的初始解选择方案、新解生成机制和当前解的改良及增加记忆功能等方法.实验结果表明,新算法传统的模拟退火算法具有更快的收敛速度和更高的稳定性.