到目前为止,有些答案建议使用邻接列表。在Java中,邻接列表如何? ......没有指针:)
我正在尝试用Java实现一个Bipartite Graph,从文件中分成两组信息。我找到了这个例子,它实际上完成了这项工作:
http://users.skynet.be/alperthereal/source_files/html/java/Bipartite.java.html
但是,我想实现我自己的版本...如果你看一下previous post of mine,你就会明白我为什么要自己做这个。
所以我必须读取一个文件,我可以从中轻松获得顶点数,但边数不是那么容易。一个示例行是“PersonA PersonB”,可以读作“PersonA说PersonB”。所以阅读这些内容......
"A says B"
"C says B"
"D says B"
"B says D"
"E says A & C"
...产生这种分组:
{A,D,C} and {B,E}.
我将如何实施这个二分图?什么是这项任务的好资源?在创建BipartiteGraph类时我应该考虑和思考什么(算法)...也许是遍历/排序算法?
答案 0 :(得分:5)
使用邻接列表实现它应该非常简单。如果它是一个无向二分图,我可能会建议使用一个关联矩阵。
因此,您可以为每个节点提供一个链接列表数组,或者某种动态分配列表的数组。它应该使添加边缘相当自然,例如在你的例子中你有一个边缘:
人A->人B
然后你去对应于Person A的数组索引并推回对应于Persona B的索引:
[人A] = B人
那么也许你会得到另一个优势
Persona A->人C
然后你的索引看起来像是:
[女主持人A] = B人,C人
作为最后一个示例,这将是示例图的邻接列表:
[A] B
[B] D
[C] B
[D] B
[E] A,C
每个索引都有一个可从该节点到达的节点列表。
“在创建BipartiteGraph类时,我应该考虑和思考什么(算法)......也许是遍历/排序算法?”
这实际上取决于你想用图表做什么......
答案 1 :(得分:3)
尝试: -
Bipartite.java
/*************************************************************************
* Compilation: javac Bipartite.java
* Dependencies: Graph.java
*
* Given a graph, find either (i) a bipartition or (ii) an odd-length cycle.
* Runs in O(E + V) time.
*
*
*************************************************************************/
/**
* The <tt>Bipartite</tt> class represents a data type for
* determining whether an undirected graph is bipartite or whether
* it has an odd-length cycle.
* The <em>isBipartite</em> operation determines whether the graph is
* bipartite. If so, the <em>color</em> operation determines a
* bipartition; if not, the <em>oddCycle</em> operation determines a
* cycle with an odd number of edges.
* <p>
* This implementation uses depth-first search.
* The constructor takes time proportional to <em>V</em> + <em>E</em>
* (in the worst case),
* where <em>V</em> is the number of vertices and <em>E</em> is the number of edges.
* Afterwards, the <em>isBipartite</em> and <em>color</em> operations
* take constant time; the <em>oddCycle</em> operation takes time proportional
* to the length of the cycle.
*/
public class Bipartite {
private boolean isBipartite; // is the graph bipartite?
private boolean[] color; // color[v] gives vertices on one side of bipartition
private boolean[] marked; // marked[v] = true if v has been visited in DFS
private int[] edgeTo; // edgeTo[v] = last edge on path to v
private Stack<Integer> cycle; // odd-length cycle
/**
* Determines whether an undirected graph is bipartite and finds either a
* bipartition or an odd-length cycle.
* @param G the graph
*/
public Bipartite(Graph G) {
isBipartite = true;
color = new boolean[G.V()];
marked = new boolean[G.V()];
edgeTo = new int[G.V()];
for (int v = 0; v < G.V(); v++) {
if (!marked[v]) {
dfs(G, v);
}
}
assert check(G);
}
private void dfs(Graph G, int v) {
marked[v] = true;
for (int w : G.adj(v)) {
// short circuit if odd-length cycle found
if (cycle != null) return;
// found uncolored vertex, so recur
if (!marked[w]) {
edgeTo[w] = v;
color[w] = !color[v];
dfs(G, w);
}
// if v-w create an odd-length cycle, find it
else if (color[w] == color[v]) {
isBipartite = false;
cycle = new Stack<Integer>();
cycle.push(w); // don't need this unless you want to include start vertex twice
for (int x = v; x != w; x = edgeTo[x]) {
cycle.push(x);
}
cycle.push(w);
}
}
}
/**
* Is the graph bipartite?
* @return <tt>true</tt> if the graph is bipartite, <tt>false</tt> otherwise
*/
public boolean isBipartite() {
return isBipartite;
}
/**
* Returns the side of the bipartite that vertex <tt>v</tt> is on.
* param v the vertex
* @return the side of the bipartition that vertex <tt>v</tt> is on; two vertices
* are in the same side of the bipartition if and only if they have the same color
* @throws UnsupportedOperationException if this method is called when the graph
* is not bipartite
*/
public boolean color(int v) {
if (!isBipartite)
throw new UnsupportedOperationException("Graph is not bipartite");
return color[v];
}
/**
* Returns an odd-length cycle if the graph is not bipartite, and
* <tt>null</tt> otherwise.
* @return an odd-length cycle (as an iterable) if the graph is not bipartite
* (and hence has an odd-length cycle), and <tt>null</tt> otherwise
*/
public Iterable<Integer> oddCycle() {
return cycle;
}
private boolean check(Graph G) {
// graph is bipartite
if (isBipartite) {
for (int v = 0; v < G.V(); v++) {
for (int w : G.adj(v)) {
if (color[v] == color[w]) {
System.err.printf("edge %d-%d with %d and %d in same side of bipartition\n", v, w, v, w);
return false;
}
}
}
}
// graph has an odd-length cycle
else {
// verify cycle
int first = -1, last = -1;
for (int v : oddCycle()) {
if (first == -1) first = v;
last = v;
}
if (first != last) {
System.err.printf("cycle begins with %d and ends with %d\n", first, last);
return false;
}
}
return true;
}
/**
* Unit tests the <tt>Bipartite</tt> data type.
*/
public static void main(String[] args) {
// create random bipartite graph with V vertices and E edges; then add F random edges
int V = Integer.parseInt(args[0]);
int E = Integer.parseInt(args[1]);
int F = Integer.parseInt(args[2]);
Graph G = new Graph(V);
int[] vertices = new int[V];
for (int i = 0; i < V; i++) vertices[i] = i;
StdRandom.shuffle(vertices);
for (int i = 0; i < E; i++) {
int v = StdRandom.uniform(V/2);
int w = StdRandom.uniform(V/2);
G.addEdge(vertices[v], vertices[V/2 + w]);
}
// add F extra edges
for (int i = 0; i < F; i++) {
int v = (int) (Math.random() * V);
int w = (int) (Math.random() * V);
G.addEdge(v, w);
}
StdOut.println(G);
Bipartite b = new Bipartite(G);
if (b.isBipartite()) {
StdOut.println("Graph is bipartite");
for (int v = 0; v < G.V(); v++) {
StdOut.println(v + ": " + b.color(v));
}
}
else {
StdOut.print("Graph has an odd-length cycle: ");
for (int x : b.oddCycle()) {
StdOut.print(x + " ");
}
StdOut.println();
}
}
}
答案 2 :(得分:1)
这是c#实现,但这个概念也可以在Java中使用。我使用Adjacency Matrix来表示图形。检查图中是否存在一个奇数周期。
如果该图中存在分区,则称为Bipartite,表示u和v其中(u union v)= Graph和(u intersection v)= null 如果你认为下图1,2,3,4,5,6,7是图G中的顶点,我们可以将左边的顶点(1,4,5,6)视为U和右边( 2,3,7)为V
现在考虑图中没有红色连接。您可以看到从u到v和v之间存在连接,因为它是一个无向图。但是在分区中没有任何关联。这就是我要使用的概念。
考虑如下图所示的图,它与上图相同,不同之处在于它更像树结构。在这种情况下,如果您可以看到备用级别上存在的节点1,3,5可以一起形成一个分区,2,4可以形成另一个分区。所以我们可以很容易地说这个图是BiPartite。如果同一级别的元素之间存在红色边缘怎么办?然后图表不是二分图。如果您可以修改BFS算法,我们就可以实现这一点。
这是代码。
int[,] BPGraph = new int[7,7]{
{0,1,0,1,0,0,0},
{1,0,1,0,1,1,0},
{0,1,0,1,0,0,1},
{1,0,1,0,1,1,0},
{0,1,0,1,0,0,1},
{0,1,0,1,0,0,1},
{0,0,1,0,1,1,0}
};
int[] BPArray = new int[7] { 0, 0, 0, 0, 0, 0, 0 };
public Boolean BiPartite()
{
Queue<int> VertexQueue = new Queue<int>();
int level = 0;
int nextlevel=0;
Boolean BPFlg = true;
VertexQueue.Enqueue(0);
while(VertexQueue.Count!=0)
{
int current = VertexQueue.Dequeue();
level = BPArray[current];
if (level == 0)
level = 1;
if (level == 2)
nextlevel=1;
else
nextlevel=2;
if(BPArray[current]==0)
BPArray[current] = level;
for (int i = 0; i < 7; i++)
{
if (BPGraph[current, i] == 1)
{
if (BPArray[i] == 0)
{
BPArray[i] = nextlevel;
VertexQueue.Enqueue(i);
}
else if (BPArray[i] == level)
{
BPFlg = false;
break;
}
}
}
if (!BPFlg)
break;
}
return BPFlg;
}
答案 3 :(得分:0)
有向图是连接节点A和B的边缘具有方向的图;如果从A到B有一条边,这并不意味着从B到A有一条边。在你的例子中,边有方向。 (B到D将是两个边,一个从B到D,一个从D到B.)
实现此目的的一种方法是以类似于链表的方式,其中节点在适当时具有彼此引用。回到您的示例,nodeA
将引用nodeB
,但不会反过来。 nodeE
会引用nodeA
和nodeC
,依此类推。你真的在创建一种(某种)数据结构,它具有节点和边缘的概念。有很多方法可以解决这个问题。
可能的Java实现将是一个名为AdjacencyList
的类,其中Map<Vertex, List<Vertex>>
包含一个顶点及其相邻顶点。然后,AdjacencyList
可以在其地图上执行操作。
至于要记住的算法和事项,请查看Wikipedia上的二分图的属性,尤其是
一个好的测试是实现一个2着色算法来确认图形确实是二分的。深度优先搜索,广度优先搜索是实施的良好练习。
答案 4 :(得分:0)
1。)选择随机节点。把它放在&#34;左边&#34;二分图的一面。
2.)选择与您在1中选择的节点相邻的所有节点,并将它们全部放在&#34;右侧&#34;侧。
3.)其余节点都属于&#34; left&#34;二分图的一面。
END