使用 Kruskal 算法计算最小生成树时的错误答案
Wrong answer while calculating Minimum Spanning Tree Using Kruskal Algorithm
我正在尝试使用 kruskal 算法求解 this MST question on spoj。我的程序似乎适用于所有测试用例,但 spoj 反复在此代码上给出 WA。
我无法在此代码上找到任何失败的测试用例。有人可以指出我做错了什么吗?
import java.io.PrintWriter;
import java.util.Arrays;
public class CSTREET {
static final int MAX = 1002;
static Node edgeList[];
static int parent[] = new int[MAX];
public static void main(String[] args) throws Exception {
Reader in = new Reader();
PrintWriter out = new PrintWriter(System.out, true);
int t = in.nextInt();
while (t-- != 0) {
int price = in.nextInt();
int vertices = in.nextInt();
int edge = in.nextInt();
int idx = 0;
edgeList = new Node[edge];
for (int i = 1; i <= vertices; i++) {
parent[i] = i;
}
while (idx < edge) {
int src = in.nextInt();
int dest = in.nextInt();
int cost = in.nextInt();
Node node = new Node(src, dest, cost);
edgeList[idx] = node;
idx++;
}
Arrays.sort(edgeList);
int edgeCount = 0;
long totalCost = 0;
idx = 0;
while (edgeCount < vertices-1 ) {
Node curEdge = edgeList[idx];
if (!checkCycle(curEdge.src, curEdge.dest)) {
edgeCount++;
totalCost += curEdge.cost;
}
idx++;
}
out.println(totalCost * price);
}
}
static boolean checkCycle(int src, int dest) {
if (findParent(src) == findParent(dest)) {
return true;
}
while (parent[dest] != parent[src]) {
parent[dest] = src;
src = parent[src];
}
return false;
}
static int findParent(int i) {
while (parent[i] != i) {
i = parent[i];
}
return i;
}
static class Node implements Comparable<Node> {
int src;
int dest;
int cost;
public Node(int src, int dest, int cost) {
this.src = src;
this.dest = dest;
this.cost = cost;
}
@Override
public int compareTo(Node o) {
return this.cost - o.cost;
}
}
}
您的 union-find 实现不正确。考虑这个例子
x -> y ( y is parent of x )
A -> B -> C
D -> E
当您调用 checkCycle( A, D)
时应该发生的是所有 5 个节点都应该转到一组,例如:
A -> B -> C
D -> E -> C
但是您的代码中发生的是:
A -> B -> C
D -> C
E
这显然是不正确的。
您可以更改 checkCycle
如下:
static boolean checkCycle(int src, int dest) {
int srcRoot = findParent(src);
int destRoot = findParent(dest);
if (srcRoot == destRoot ) {
return true;
}
parent[destRoot] = srcRoot;
return false;
}
我强烈建议你阅读关于 Disjoint-set 的维基百科文章并实现路径压缩版本,这会提高复杂度。
我正在尝试使用 kruskal 算法求解 this MST question on spoj。我的程序似乎适用于所有测试用例,但 spoj 反复在此代码上给出 WA。
我无法在此代码上找到任何失败的测试用例。有人可以指出我做错了什么吗?
import java.io.PrintWriter;
import java.util.Arrays;
public class CSTREET {
static final int MAX = 1002;
static Node edgeList[];
static int parent[] = new int[MAX];
public static void main(String[] args) throws Exception {
Reader in = new Reader();
PrintWriter out = new PrintWriter(System.out, true);
int t = in.nextInt();
while (t-- != 0) {
int price = in.nextInt();
int vertices = in.nextInt();
int edge = in.nextInt();
int idx = 0;
edgeList = new Node[edge];
for (int i = 1; i <= vertices; i++) {
parent[i] = i;
}
while (idx < edge) {
int src = in.nextInt();
int dest = in.nextInt();
int cost = in.nextInt();
Node node = new Node(src, dest, cost);
edgeList[idx] = node;
idx++;
}
Arrays.sort(edgeList);
int edgeCount = 0;
long totalCost = 0;
idx = 0;
while (edgeCount < vertices-1 ) {
Node curEdge = edgeList[idx];
if (!checkCycle(curEdge.src, curEdge.dest)) {
edgeCount++;
totalCost += curEdge.cost;
}
idx++;
}
out.println(totalCost * price);
}
}
static boolean checkCycle(int src, int dest) {
if (findParent(src) == findParent(dest)) {
return true;
}
while (parent[dest] != parent[src]) {
parent[dest] = src;
src = parent[src];
}
return false;
}
static int findParent(int i) {
while (parent[i] != i) {
i = parent[i];
}
return i;
}
static class Node implements Comparable<Node> {
int src;
int dest;
int cost;
public Node(int src, int dest, int cost) {
this.src = src;
this.dest = dest;
this.cost = cost;
}
@Override
public int compareTo(Node o) {
return this.cost - o.cost;
}
}
}
您的 union-find 实现不正确。考虑这个例子
x -> y ( y is parent of x )
A -> B -> C
D -> E
当您调用 checkCycle( A, D)
时应该发生的是所有 5 个节点都应该转到一组,例如:
A -> B -> C
D -> E -> C
但是您的代码中发生的是:
A -> B -> C
D -> C
E
这显然是不正确的。
您可以更改 checkCycle
如下:
static boolean checkCycle(int src, int dest) {
int srcRoot = findParent(src);
int destRoot = findParent(dest);
if (srcRoot == destRoot ) {
return true;
}
parent[destRoot] = srcRoot;
return false;
}
我强烈建议你阅读关于 Disjoint-set 的维基百科文章并实现路径压缩版本,这会提高复杂度。