[人工智能AI]之搜索-A*算法

[人工智能AI]之搜索-A*算法

1. 概述

这是一种经典的寻路算法；
简而言之，重点在于估算距离。

---

问题描述：

假设起点a, 终点b, 当前结点为n
F(n) = G(n) + H(n)
G(n)表示起点到n实际距离；
H(n)表示n到终点预估距离；
在所有结点中选择F值最小的可行节点走下去即可。

---

缺陷：

1. 显然，A*算法的效率很大程度取决于H(n)的计算，即对未来距离的预估；
2. 未完待续…

---

2. A*算法细节

2.1 对于H(n)估算函数的选择

曼哈顿距离：横纵坐标差之和 (推荐使用)
欧氏距离：即两点间的实际距离
ps: 这两种估算方式，会导致无法避开障碍物。

---

2.2 A*算法实现

"""
A*寻路算法, 评估函数采用曼哈顿距离
"""
path = [[0, 0, 0, 0, 0, 0, 0, 0, 0],
        [0, 0, 0, 3, 0, 0, 0, 0, 0],
        [0, 0, 0, 3, 0, 0, 2, 0, 0],
        [0, 0, 0, 3, 0, 0, 0, 0, 0],
        [0, 4, 0, 3, 0, 0, 0, 0, 0],
        [0, 0, 0, 3, 0, 0, 0, 0, 0],
        [0, 0, 0, 3, 0, 0, 0, 0, 0],
        [0, 0, 0, 0, 0, 0, 0, 0, 0],
        ]

m = len(path)
n = len(path[0])

visit = 0

a = [4, 1]
b = [2, 6]

def showNode():
    for i in path:
        for j in i:
            print(j, end='')
        print()
    print()

def getNode(x):
    return path[x[0]][x[1]]

def inSide(x, y):
    return x>=0 and x<m and y>=0 and y<n

def isNode(x, y):
    return path[x][y] == 0 or path[x][y] == 2

def getNeigh(x, y):
    neigh = []
    if inSide(x-1, y) and isNode(x-1, y):
        neigh.append([x-1, y])
    if inSide(x+1, y) and isNode(x+1,y):
        neigh.append([x+1, y])
    if inSide(x, y-1) and isNode(x,y-1):
        neigh.append([x, y-1])
    if inSide(x, y+1) and isNode(x,y+1):
        neigh.append([x, y+1])
    return neigh

def testNeigh(neigh):
    for i in range(m):
        for j in range(n):
            if [i, j] in neigh:
                print(' ', end='')
            else:
                print(path[i][j], end='')
        print()

def calDist(A, B):
    # 计算两点的曼哈顿距离
    return abs(A[0]-B[0]) + abs(A[1]-B[1])

def update(dist, neigh):
    minDist = 0x7fffffff
    minXY = 0
    for node in neigh: # like [3, 4]
        dist[node[0]][node[1]] = visit + calDist(node, b)
        if dist[node[0]][node[1]] < minDist:
            minDist = dist[node[0]][node[1]]
            minXY = node
    return minXY

def findPath(curX, curY):
    global path, visit
    showNode()
    if curX == b[0] and curY == b[1]:
        print('Reach the End!')
        return
    dist = [[-1 for i in range(n)] for i in range(m)]
    neigh = getNeigh(curX, curY)
    XY = update(dist, neigh)
    visit += 1
    curX, curY = XY[0], XY[1]
    path[curX][curY] = ' '
    findPath(curX, curY)
    # 更新

findPath(a[0], a[1])

演示结果(节选部分)如下图, 起点为4, 终点为2, 障碍物为3, 其他0表示可行点