图论在数据结构中是非常有趣而复杂的,作为 Web 码农的我,在实际开发中一直没有找到它的使用场景,不像树那样的频繁使用,不过还是准备仔细的把图论全部过一遍。
图中有一个好玩的东西叫做生成树,就是用边来把所有的顶点联通起来,前提条件是最后形成的联通图中不能存在回路,所以就形成这样一个推理:假设图中的顶点有 n 个,则生成树的边有 n-1 条,多一条会存在回路,少一路则不能把所有顶点联通起来,如果非要在图中加上权重,则生成树中权重最小的叫做最小生成树。
对于上面这个带权无向图来说,它的生成树有多个,同样最小生成树也有多个,因为我们比的是权重的大小。
求最小生成树的算法有很多,常用的是 Prim 算法和 Kruskal 算法,为了保证单一职责,我把 Kruskal 算法放到下一篇,那么 Prim 算法的思想是什么呢?很简单,贪心思想。
如上图:现有集合 M={A,B,C,D,E,F},再设集合 N={}。
最后 M 集合为{}时,生成树就构建完毕了,是不是非常的简单,这种贪心做法我想大家都能想得到,如果算法配合一个好的数据结构,就会如虎添翼。
图的存储有很多方式,邻接矩阵,邻接表,十字链表等等,当然都有自己的适合场景,下面用邻接矩阵来玩玩,邻接矩阵需要采用两个数组,
①. 保存顶点信息的一维数组,
②. 保存边信息的二维数组。
public class Graph
{
///
/// 顶点个数
///
public char[] vertexs;
///
/// 边的条数
///
public int[,] edges;
///
/// 顶点个数
///
public int vertexsNum;
///
/// 边的个数
///
public int edgesNum;
}
矩阵构建很简单,这里把上图中的顶点和权的信息保存在矩阵中。
#region 矩阵的构建
///
/// 矩阵的构建
///
public void Build()
{
//顶点数
graph.vertexsNum = 6;
//边数
graph.edgesNum = 8;
graph.vertexs = new char[graph.vertexsNum];
graph.edges = new int[graph.vertexsNum, graph.vertexsNum];
//构建二维数组
for (int i = 0; i < graph.vertexsNum; i++)
{
//顶点
graph.vertexs[i] = (char)(i + 65);
for (int j = 0; j < graph.vertexsNum; j++)
{
graph.edges[i, j] = int.MaxValue;
}
}
graph.edges[0, 1] = graph.edges[1, 0] = 80;
graph.edges[0, 3] = graph.edges[3, 0] = 100;
graph.edges[0, 5] = graph.edges[5, 0] = 20;
graph.edges[1, 2] = graph.edges[2, 1] = 90;
graph.edges[2, 5] = graph.edges[5, 2] = 70;
graph.edges[3, 2] = graph.edges[2, 3] = 100;
graph.edges[4, 5] = graph.edges[5, 4] = 40;
graph.edges[3, 4] = graph.edges[4, 3] = 60;
graph.edges[2, 3] = graph.edges[3, 2] = 10;
}
#endregion
要玩 Prim,我们需要两个字典。
①:保存当前节点的字典,其中包含该节点的起始边和终边以及权值,用 weight=-1 来记录当前节点已经访问过,用 weight=int.MaxValue 表示两节点没有边。
②:输出节点的字典,存放的就是我们的 N 集合。
当然这个复杂度玩高了,为 O(N2),寻找 N 集合的邻边最小权值时,我们可以玩玩 AVL 或者优先队列来降低复杂度。
#region prim算法
///
/// prim算法
///
public Dictionary<char, Edge> Prim()
{
Dictionary<char, Edge> dic = new Dictionary<char, Edge>();
//统计结果
Dictionary<char, Edge> outputDic = new Dictionary<char, Edge>();
//weight=MaxValue:标识没有边
for (int i = 0; i < graph.vertexsNum; i++)
{
//起始边
var startEdge = (char)(i + 65);
dic.Add(startEdge, new Edge() { weight = int.MaxValue });
}
//取字符的开始位置
var index = 65;
//取当前要使用的字符
var start = (char)(index);
for (int i = 0; i < graph.vertexsNum; i++)
{
//标记开始边已使用过
dic[start].weight = -1;
for (int j = 1; j < graph.vertexsNum; j++)
{
//获取当前 c 的 邻边
var end = (char)(j + index);
//取当前字符的权重
var weight = graph.edges[(int)(start) - index, j];
if (weight < dic[end].weight)
{
dic[end] = new Edge()
{
weight = weight,
startEdge = start,
endEdge = end
};
}
}
var min = int.MaxValue;
char minkey = ' ';
foreach (var key in dic.Keys)
{
//取当前 最小的 key(使用过的除外)
if (min > dic[key].weight && dic[key].weight != -1)
{
min = dic[key].weight;
minkey = key;
}
}
start = minkey;
//边为顶点减去1
if (outputDic.Count < graph.vertexsNum - 1 && !outputDic.ContainsKey(minkey))
{
outputDic.Add(minkey, new Edge()
{
weight = dic[minkey].weight,
startEdge = dic[minkey].startEdge,
endEdge = dic[minkey].endEdge
});
}
}
return outputDic;
}
#endregion
public static void Main()
{
MatrixGraph martix = new MatrixGraph();
martix.Build();
var dic = martix.Prim();
Console.WriteLine("最小生成树为:");
foreach (var key in dic.Keys)
{
Console.WriteLine("({0},{1})({2})", dic[key].startEdge, dic[key].endEdge, dic[key].weight);
}
Console.Read();
}
using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Diagnostics;
using System.Threading;
using System.IO;
using SupportCenter.Test.ServiceReference2;
using System.Threading.Tasks;
namespace ConsoleApplication2
{
public class Program
{
public static void Main()
{
MatrixGraph martix = new MatrixGraph();
martix.Build();
var dic = martix.Prim();
Console.WriteLine("最小生成树为:");
foreach (var key in dic.Keys)
{
Console.WriteLine("({0},{1})({2})", dic[key].startEdge, dic[key].endEdge, dic[key].weight);
}
Console.Read();
}
}
///
/// 定义矩阵节点
///
public class MatrixGraph
{
Graph graph = new Graph();
public class Graph
{
///
/// 顶点个数
///
public char[] vertexs;
///
/// 边的条数
///
public int[,] edges;
///
/// 顶点个数
///
public int vertexsNum;
///
/// 边的个数
///
public int edgesNum;
}
#region 矩阵的构建
///
/// 矩阵的构建
///
public void Build()
{
//顶点数
graph.vertexsNum = 6;
//边数
graph.edgesNum = 8;
graph.vertexs = new char[graph.vertexsNum];
graph.edges = new int[graph.vertexsNum, graph.vertexsNum];
//构建二维数组
for (int i = 0; i < graph.vertexsNum; i++)
{
//顶点
graph.vertexs[i] = (char)(i + 65);
for (int j = 0; j < graph.vertexsNum; j++)
{
graph.edges[i, j] = int.MaxValue;
}
}
graph.edges[0, 1] = graph.edges[1, 0] = 80;
graph.edges[0, 3] = graph.edges[3, 0] = 100;
graph.edges[0, 5] = graph.edges[5, 0] = 20;
graph.edges[1, 2] = graph.edges[2, 1] = 90;
graph.edges[2, 5] = graph.edges[5, 2] = 70;
graph.edges[3, 2] = graph.edges[2, 3] = 100;
graph.edges[4, 5] = graph.edges[5, 4] = 40;
graph.edges[3, 4] = graph.edges[4, 3] = 60;
graph.edges[2, 3] = graph.edges[3, 2] = 10;
}
#endregion
#region 边的信息
///
/// 边的信息
///
public class Edge
{
//开始边
public char startEdge;
//结束边
public char endEdge;
//权重
public int weight;
}
#endregion
#region prim算法
///
/// prim算法
///
public Dictionary<char, Edge> Prim()
{
Dictionary<char, Edge> dic = new Dictionary<char, Edge>();
//统计结果
Dictionary<char, Edge> outputDic = new Dictionary<char, Edge>();
//weight=MaxValue:标识没有边
for (int i = 0; i < graph.vertexsNum; i++)
{
//起始边
var startEdge = (char)(i + 65);
dic.Add(startEdge, new Edge() { weight = int.MaxValue });
}
//取字符的开始位置
var index = 65;
//取当前要使用的字符
var start = (char)(index);
for (int i = 0; i < graph.vertexsNum; i++)
{
//标记开始边已使用过
dic[start].weight = -1;
for (int j = 1; j < graph.vertexsNum; j++)
{
//获取当前 c 的 邻边
var end = (char)(j + index);
//取当前字符的权重
var weight = graph.edges[(int)(start) - index, j];
if (weight < dic[end].weight)
{
dic[end] = new Edge()
{
weight = weight,
startEdge = start,
endEdge = end
};
}
}
var min = int.MaxValue;
char minkey = ' ';
foreach (var key in dic.Keys)
{
//取当前 最小的 key(使用过的除外)
if (min > dic[key].weight && dic[key].weight != -1)
{
min = dic[key].weight;
minkey = key;
}
}
start = minkey;
//边为顶点减去1
if (outputDic.Count < graph.vertexsNum - 1 && !outputDic.ContainsKey(minkey))
{
outputDic.Add(minkey, new Edge()
{
weight = dic[minkey].weight,
startEdge = dic[minkey].startEdge,
endEdge = dic[minkey].endEdge
});
}
}
return outputDic;
}
#endregion
}
}