如何使用邻接矩阵找到未指向,未加权图的最重的路径(图的M(d))的重量
图像中最长的路径是最长的路径,但LM是最重的 该图的最重的路径是连接到其最边缘的路径,对于此当前情况,最重的路径是LM,具有19个边缘:LK,KJ,JI,IH,IH,HG,HG,GF,GF,FE,FM,FM,FM,MN,MO,MO,MO,MO,MO,MO, MP,MQ,MR,MS,MT,MU,MV,MV,MW,MX。 我正在尝试创建一个程序,该程序将检查输入的矩阵是否代表树图,如果确实如此,则返回图形最重的路径上的边数。换句话说,它将返回图的m(d)。 对于当前情况,输出将是连接到LM路径的边数,即19岁。
但是我找不到这样做的方法。在这里,我创建了一个函数,以找到至少1秒以将其设置为根,但不知道该如何进行。
let matrix = [
[0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
//find row with least 1s
function findRowWithLeastOnes(matrix_data) {
let counter = 0;
let row = 0;
for (let i = 0; i < matrix_data.length; i++) {
for (let j = 0; j < matrix_data[i].length; j++) {
if (matrix_data[i][j] === 1) {
counter++;
}
}
if (counter < matrix_data.length) {
row = i;
console.log(row);
return row;
}
}
}
findRowWithLeastOnes(matrix);
In the image A-L is the longest path, but L-M is the heaviest
The heaviest path of the graph is the path with the most edges connected to it, for this current case the heaviest path is L-M with 19 edges: L-K, K-J, J-I, I-H, H-G, G-F, F-E, F-M, M-N, M-O, M-P, M-Q, M-R, M-S, M-T, M-U, M-V, M-W, M-X.
I am trying to create a program that will check that the inputted matrix represents a tree graph, and if it does, that returns the number of edges on the heaviest path of the graph. In other words it will return the M(d) of the graph.
For this current case the output will be the number of edges connected to L-M path which is 19.
But I can not find a way to do that. Here I created a function to find the row with the least 1s to set it as a root, but don't know how to proceed.
let matrix = [
[0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
//find row with least 1s
function findRowWithLeastOnes(matrix_data) {
let counter = 0;
let row = 0;
for (let i = 0; i < matrix_data.length; i++) {
for (let j = 0; j < matrix_data[i].length; j++) {
if (matrix_data[i][j] === 1) {
counter++;
}
}
if (counter < matrix_data.length) {
row = i;
console.log(row);
return row;
}
}
}
findRowWithLeastOnes(matrix);
如果你对这篇内容有疑问,欢迎到本站社区发帖提问 参与讨论,获取更多帮助,或者扫码二维码加入 Web 技术交流群。

绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱,如果其他用户或者作者回复了您的评论,将不能在第一时间通知您!
发布评论
评论(1)
我们可以使用以下算法:
将矩阵变成邻接列表,以避免始终扫描零,并可以轻松访问节点的度,从而有助于重量。
选择以(例如0)开始的任何节点,然后从中找到最重的路径。我们可以对此进行深度优先搜索(例如,使用递归)。从发现的所有路径中,保持最重。为避免边缘来回穿越,请记住以前的节点在遍历中是什么:不应重新审视它。如果有一个先前的节点,我们应该避免在权重总和中对遍历边缘进行两次计数。
以该路径中的最后一个节点:这是远离我们开始节点的节点。该节点将是我们要寻找的最重的路径的终点。
第二次使用相同的算法来找到该节点最重的路径。这是我们需要的路径。
如果您的问题还涉及确定图是否是树,那么您本质上需要检查没有周期(深度第一个遍历可以再次使用),并且可以从一个节点中达到所有其他节点(首先是深度首先可以服务深度)。也许您还想检查图表是否没有方向,即矩阵沿主角镜像镜像。
We can use the following algorithm:
Turn the matrix into an adjacency list so to avoid scanning zeroes all the time, and to have easy access to the degree of a node, which contributes to the weight.
Choose any node to start with (e.g. 0) and find the heaviest path from it. We can use a depth-first search for that (for instance, using recursion). From all paths that are found, keep the heaviest. To avoid an edge being traversed back and forth, remember what the previous node was in the traversal: it should not be revisited. If there was a previous node, we should avoid counting the traversed edge twice in the sum of weights.
Take the last node in that path: this is the node that is the furthest away from our starting node in terms of weight. This node will be an end-point of the heaviest path we are looking for.
Use the same algorithm a second time to find the heaviest path from that node. This is the path we need.
If your question is also about determining whether a graph is a tree, you would essentially need to check there are no cycles (a depth first traversal can be used again) and that from one node all others can be reached (again depth first can serve). Maybe you'd also want to check that the graph is undirected, i.e. that the matrix is mirrored along the main diagonal.