该文就基于线性单调多边形模型的二维mesh中是否存在曼哈顿路径,分别给出了最小路由和启发式路由两种解决方案,最终的实验结果表明,对于给定的错误率的不同规模的二维mesh网络,在随机产生错误结点时,两者的平均路径长度与曼哈顿路径长度偏离很小.另外,多计算机系统中某对特定源/目的结点间的路由被应用的次数极少,因而在路径长度和时间开销间寻找折衷比穷尽搜索绝对最短路径更有实际意义.“,”Based on the existence of Manhattan routes in presence of rectilinear-monotone polygonal fault blocks in 2-D mesh,minimal routing and heuristic routing schemes are proposed.The final experimental results indicate that for giving fault rates in 2 -D meshes of various sizes,minimal and heuristic route lengths deviate from Manhattan route length only by a small distance when fault nodes are randomly generated.Furthermore,in a multiprocessor computer system,a route for a specific source/destination may only be used a few times,so it is more significant to seek for a trade-off between route length and time spending than exhaustive search for a shortest route.