我们尝试用禁忌搜索(tabu search）算法来解决旅行商问题（TSP），程序可以选择随机生成点的位置坐标，也可以从圆上均匀取点（比较容易判断结果的好坏）。

我们采用下面的一些规则来运行程序：

特赦规则：如果新解被禁，但是目标函数值比当前最优解的目标函数值小于某个定值时，接受新解。

终止规则：1、如果在给定的迭代次数内，最优解的目标函数值保持不变终止2、迭代次数达到给定的最大值时终止。

候选解的产生：本实验中我采取的是随机产生候选解的方法

一些参数的选取：

在每一步的迭代中，保留bl n个候选解（也可以产生n(n-1)/2），禁忌长度取tl sqrt(n),最大迭代步数设为maxl 500

运行程序有如下结果：

在半径为5的圆上均匀取n 50个点（能够直观地观察解的收敛性）-We try Tabu search (tabu search) algorithm to solve the traveling salesman problem (TSP), the program can　the location coordinates of randomly generated points can even take some (relatively easy to determine the result is good or bad)　the circle. We use the following rules to run the program: Amnesty Rule: If the new solution is disabled, but the objective function value is less than a given value than the current optimal solution of the objective function, the acceptance of the new solution. Termination rules: 1. If, within a given number of iterations, the optimal solution of the objective function value remains unchanged termination 2, termination achieve a given maximum number of iterations. Candidate Generation Solutions: The experiment I took a random method of generating a candidate solution of some of the selected parameters: at each step of the iteration, bl　 n candidate solution retention (can also produce n (n-1)/2) taboos length take tl　 sqrt (n), the maximum number of