用的逐步定线方法，在解决这个问题中并不是一个很有效的方法，我想最好的方法为Dijkstra，但Dijkstra编起来比较简单（已有源程序，只须改变权矩阵），没有再进行练习。

程序编制的比一些比赛论文中的要精细一些，但结果不是很好。-The progressive alignment method used in solving this problem is not a very effective way, I think the best way to Dijkstra, but it is relatively simple Dijkstra Code (has source code, just change the weight matrix), no further to practice. Programming than some games should be fine in a number of papers, but the results are not very good.