模拟退火算法中冷却调度选取方法的研究

Research on Cooling Schedule Selecting Method in Simulated Annealing Algorithm

模拟退火算法是解决组合优化问题的有效方法，冷却调度是它的关键部分．该文给出了初始温度、降温策略、Ｍａｒｋｏｖ链长度以及停止准则的选取方法，尤其是在停止准则方面，借鉴禁忌搜索的思想，提出了一种新的模拟退火算法停止准则．以上方法在中国３１城市旅行商问题以及抖动模式问题中得到了有效应用．

Cooling schedule is the key to simulated annealing algorithm that is effective means to solvs combination optimal problem. In this paper. some selecting criteria are given about cooling schedule including initial temperature. temperature dropping strategy. Markov chain length and stopping criterion, especiallv about stopping criterion of simulated annealing algorithm. a new criterion is presented similar to the one in tabu search. The above conclusions are successfully applied in the traveling salesman problem of 31 cities in China and in the dithering pattern problem.