A*算法介绍及C#实现
参考:
http://en.wikipedia.org/wiki/A*_search_algorithm
介绍:
AStar算法是最短路径搜索的一种,类属于启发式搜索. :), 照搬。。。。
这东西还是有点麻烦的,请耐点心,有几个核心步骤需要解释清楚
公式:
F = G + H
G: 从现在访问这点到开始点的消耗
H: 从现在这点到结束点的估计消耗
F: 做判定访问优先级的依据
数据结构:
1. Open, Close队列
Open里保存已经打开但是并未访问的元素
2. FScore, GScore, HScore 列表
保存该元素所需要的消耗
3. ComeFrom 每个点的父亲关系
操作概要解释:
每次都找他周围看起来最容易达到终点的元素,如果周围的点G值比他小,则先走G值小的位置,再判定该位置F值最小的点,不停循环,直到找到终点为止。
该算法需要先走遍G值小的部分,而后,才会走F值的判断,这样先一直往前,如果不行,则找看起来最有可能快到终点的元素,这样ComeFrom里会有一些非最短路径之外的位置,不过根据父子关系也可以很容易的得到路径
代码实现(C#版本)
1: using System;
2: using System.Collections;
3: using System.Collections.Generic;
4:
5: namespace Roger.Testing
6: {
7: public class Testing
8: {
9: public static void Main()
10: {
11: List<bool> map = new List<bool>
12: {
13: true, true, true, false, false,
14: false, true, false, true,false,
15: false, false, false, true,false,
16: false,true,false, false, false,
17: false, false, false, true, false
18: };
19: int width = 5;
20: int start = 10;
21: int end = 24;
22:
23: if (map[start] || map[end]) {
24: Console.Write("起始点或者结束点为不合法障碍物,!");
25: Console.Read();
26: return;
27: }
28:
29: AStarSearch finder = new AStarSearch(map, width);
30: Dictionary<int, int> path = finder.FindPath(start, end);
31:
32: if (path == null)
33: {
34: Console.WriteLine("没有找到对应的路径"); Console.Read(); return;
35: }
36:
37: // 得到最后一个Key的值
38: int key = end;
39: List<int> link = new List<int>() { end };
40: while (path[key] != start)
41: {
42: key = path[key];
43: link.Add(key);
44: }
45: link.Add(start);
46:
47: for (int i = 0; i < map.Count; i++ )
48: {
49: if (i % width == 0) Console.WriteLine();
50: if (map[i]) Console.Write("#");
51: else if (link.Contains(i)) Console.Write("1");
52: else Console.Write("*");
53:
54: Console.Write(" ");
55: }
56:
57: Console.Read();
58: }
59: }
60:
61: /// <summary>
62: /// A* 最短路径搜索
63: /// </summary>
64: public class AStarSearch
65: {
66: private List<bool> Map;
67: private int Width;
68: private int Height;
69: private bool IsValidPosition(int start)
70: {
71: return start >= 0 && start <= Map.Count;
72: }
73: public AStarSearch(List<bool> map, int Width)
74: {
75: if (map == null) throw new ArgumentNullException();
76:
77: Map = map;
78: this.Width = Width;
79: this.Height = Map.Count / Width;
80: }
81:
82: private Queue<int> Open;
83: private Queue<int> Close;
84:
85: public Dictionary<int, int> FindPath(int start, int end)
86: {
87: if (!IsValidPosition(start) || !IsValidPosition(end)) throw new ArgumentOutOfRangeException();
88: this.Start = start; this.End = end;
89: Open = new Queue<int>();
90: Close = new Queue<int>();
91: GScore = new Dictionary<int, int>();
92: FScore = new Dictionary<int, int>();
93: ComeFrom = new Dictionary<int, int>();
94:
95: // 将开始节点入队列
96: Open.Enqueue(start);
97:
98: int x = start;
99: while (Open.Count > 0)
100: {
101: x = GetLowestF();
102: if (x == End)
103: {
104: // Trace From
105: return ComeFrom;
106: }
107:
108: Open.Dequeue();
109: Close.Enqueue(x);
110:
111: foreach (int y in GetNodesAround(x))
112: {
113: if (Close.Contains(y))
114: {
115: continue;
116: }
117:
118: int newGValue = GetCost(x) + GetDistance(x, y);
119: bool newIsBetter = false;
120:
121: if (!Open.Contains(y))
122: {
123: Open.Enqueue(y);
124: newIsBetter = true;
125: }
126: else if (newGValue < GScore[y])
127: newIsBetter = true;
128:
129:
130: if(newIsBetter)
131: {
132: if (ComeFrom.ContainsKey(y))
133: ComeFrom[y] = x;
134: else
135: ComeFrom.Add(y, x);
136:
137: GScore[y] = newGValue;
138: FScore[y] = GScore[y] + GetHeuristic(y);
139: }
140:
141: }
142: }
143:
144: return null;
145: }
146:
147: private int Start;
148: private int End;
149:
150:
151: private IList<int> GetNodesAround(int pos)
152: {
153: List<int> list = new List<int>(4);
154: int x = pos % Width; int y = pos / Width;
155: // Left
156: if (x > 0 && !Map[x - 1 + y * Width]) list.Add(x - 1 + y * Width);
157: // Up
158: if (y > 0 && !Map[x + (y - 1) * Width]) list.Add(x + (y - 1) * Width);
159: // Right
160: if (x < Width-1 && !Map[x + 1 + y * Width]) list.Add(x + 1 + y * Width);
161: // Down
162: if (y < Height-1 && !Map[x + (y + 1) * Width]) list.Add(x + (y + 1) * Width);
163:
164: return list;
165: }
166:
167: private int GetCost(int current)
168: {
169: // UNDONE
170: int xDistance = (int)Math.Abs(Start % Width - current % Width);
171: int yDistance = (int)Math.Abs(Start / Width - current / Width);
172:
173: if (!GScore.ContainsKey(current))
174: GScore.Add(current ,(xDistance + yDistance) * 10);
175:
176: return GScore[current];
177: }
178:
179: private int GetLowestF()
180: {
181: int temp = GetFScore(Open.Peek());
182: int lowest = Open.Peek();
183:
184: foreach (int i in Open)
185: {
186: if (temp > GetFScore(i))
187: {
188: temp = GetFScore(i);
189: lowest = i;
190: }
191: }
192:
193: return lowest;
194: }
195:
196: private int GetFScore(int pos)
197: {
198: if (!FScore.ContainsKey(pos))
199: FScore.Add(pos, GetCost(pos) + GetHeuristic(pos));
200:
201: return FScore[pos];
202: }
203:
204: private Dictionary<int, int> GScore;
205: private Dictionary<int, int> FScore;
206: private Dictionary<int, int> ComeFrom;
207:
208: // 得到预估的距离
209: private int GetHeuristic(int current)
210: {
211: return GetDistance(current, End);
212: }
213:
214: private int GetDistance(int x, int y)
215: {
216: int xDistance = (int)Math.Abs(y % Width - x % Width);
217: int yDistance = (int)Math.Abs(y / Width - x / Width);
218:
219: return 10 * (xDistance + yDistance);
220:
221: }
222: }
223: }
有同志说,用队列的方式效率低下,的确如此,用二项堆这样的方法会快一些,毕竟开始就是排好序的,奇怪的是,我自己写一个堆排序结果却不好,郁闷的紧。
另外用SortDictionary做了测试,最小值做的Value, 没有做Key,因为Key只能用标记点的位置代表
代码包:AStar_Source
Update: 使用二项堆与SortedDictionary混合排序确实会提高效率,至于原因neoragex2002已经讲清楚了。
BinaryHeap适用于最小或者最大的排序,而SortedDictionary是平衡二叉树(红黑树),运行效率会高一些,再次感谢neoragex2002 :)
输出结果如下: