gpt4 book ai didi

java - 使用DFS回溯算法生成迷宫的问题

转载 作者:行者123 更新时间:2023-12-01 18:15:57 26 4
gpt4 key购买 nike

我正在尝试实现 DFS 回溯算法,该算法涉及利用维基百科上找到的堆栈(而不是递归算法)。我试图生成一个由 0 和 1 组成的迷宫,其中 1 代表一堵墙,0 代表一条可用路径。对于迷宫中任何不是墙的给定空间,必须始终存在一条可以从任何其他非墙单元到达的有效路径。

我从一个大小为 maze[15][20] 的二维数组的迷宫开始,并按照算法标记需要标记为已访问的单元格。最初,所有单元格(不包括外边界)都被标记为“未访问”。所有单元格都初始化为值“1”,并且期望算法将在整个迷宫中挖掘独特的路径。

算法链接(利用堆栈的递归回溯器第二次实现):

https://en.wikipedia.org/wiki/Maze_generation_algorithm

我写的代码:

public void innerMaze() {
List<Coordinate> listOfCoordinates = new ArrayList<>();
List<Coordinate> coordinatesToBeRemoved = new ArrayList<>();
Stack<Coordinate> DFS_Stack = new Stack();

Coordinate initialCell = new Coordinate(1, 1);
checkIfVisited.put(initialCell, true);
DFS_Stack.push(initialCell);

int randomInteger = 0;
int cx = 0;
int cy = 0;
int gx = 0;
int gy = 0;

while (!DFS_Stack.empty()) {
Coordinate currentCoordinate = DFS_Stack.pop();
cx = currentCoordinate.getX();
cy = currentCoordinate.getY();

if ((cx - 2) >= 1) {
Coordinate up = findCoordinate((cx - 2), cy);
up.setDirection('N');
listOfCoordinates.add(up);

}
if ((cx + 2) <= MAX_ROW) {
Coordinate down = findCoordinate((cx + 2), cy);
down.setDirection('S');
listOfCoordinates.add(down);
}
if ((cy - 2) >= 1) {
Coordinate left = findCoordinate(cx, (cy - 2));
left.setDirection('W');
listOfCoordinates.add(left);
}
if ((cy + 2) <= MAX_COL) {
Coordinate right = findCoordinate(cx, (cy + 2));
right.setDirection('E');
listOfCoordinates.add(right);
}
for (Coordinate s : listOfCoordinates) {
if (checkIfVisited.get(s) == true) {
coordinatesToBeRemoved.add(s);
}
}
listOfCoordinates.removeAll(coordinatesToBeRemoved);

if (!listOfCoordinates.isEmpty()) {
DFS_Stack.push(currentCoordinate);
randomInteger = ThreadLocalRandom.current().nextInt(0, listOfCoordinates.size());
Coordinate temp = listOfCoordinates.get(randomInteger);
char direction = temp.getDirection();
Coordinate newWall;

if (direction == 'N') {
newWall = findCoordinate((cx - 1), cy);
} else if (direction == 'S') {
newWall = findCoordinate((cx + 1), cy);
} else if (direction == 'W') {
newWall = findCoordinate(cx, (cy - 1));
} else {
newWall = findCoordinate(cx, (cy + 1));
}
System.out.println(newWall);
gx = newWall.getX();
gy = newWall.getY();
completeMaze[gx][gy] = 0;
checkIfVisited.put(temp, true);
checkIfVisited.put(newWall, true);
listOfCoordinates.clear();
DFS_Stack.push(temp);
}
}
}

在我当前的实现中,每个单元格要么代表一堵墙,要么代表一条路径,因此我稍微改变了算法,删除两个单元格之间的墙就变成了删除两个单元格之外的墙,将中间的一个更改为一堵墙。我的输出示例如下:

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 1 0 1 1 1 0 1 0 1 0 1 0 1 1 1 0 1 0 1
1 0 1 1 1 0 1 1 1 0 1 1 1 0 1 1 1 1 1 0
1 1 1 1 0 1 0 1 1 1 1 1 1 1 0 1 0 1 0 1
1 0 1 1 1 0 1 1 1 0 1 0 1 1 1 1 1 1 1 0
1 1 0 1 1 1 1 1 0 1 1 1 0 1 0 1 1 1 0 1
1 1 1 0 1 0 1 0 1 1 1 1 1 1 1 0 1 0 1 1
1 1 1 1 1 1 1 1 0 1 1 1 0 1 1 1 1 1 1 1
1 0 1 0 1 0 1 1 1 0 1 0 1 0 1 0 1 0 1 0
1 1 0 1 1 1 1 1 0 1 1 1 1 1 0 1 1 1 1 1
1 0 1 1 1 1 1 0 1 1 1 0 1 1 1 0 1 0 1 0
1 1 0 1 1 1 0 1 1 1 1 1 0 1 1 1 1 1 1 1
1 1 1 0 1 0 1 1 1 0 1 0 1 0 1 1 1 0 1 0
1 1 0 1 0 1 1 1 0 1 0 1 1 1 0 1 0 1 0 1
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

乍一看,二维数组索引[1][1]被墙壁包围,因此它是一个无法到达的区域。通过多次执行,这也非常一致。

如有任何帮助,我们将不胜感激。

最佳答案

我强烈建议使用不同的算法。我在编码迷宫生成算法方面拥有丰富的经验,并且递归回溯总是会导致非常长的路径,几乎没有任何选择。我在你的代码中看不到这个问题,但这是我自己的递归回溯的 Java 实现(如果代码不好,抱歉,它已经有几年了)。

public class Maze {

private int sizeX, sizeY;
private Cell[][] cells;
private List<Cell> backtracker = new ArrayList<>();

public Maze(int sizeX, int sizeY) {
this.sizeX = sizeX;
this.sizeY = sizeY;
this.cells = new Cell[sizeX][sizeY];
for (int i = 0; i < sizeX; i++) {
for (int j = 0; j < sizeY; j++)
cells[i][j] = new Cell(this, i, j);
}
generate();
}

private void generate() {
Cell current = cells[0][0];
current.visit();
boolean hasUnvisited = hasUnvisited();
while (hasUnvisited) {
current = current.pickNext();
if (!current.isVisited()) {
current.visit();
backtracker.add(current);
hasUnvisited = hasUnvisited();
}
}
}

private boolean hasUnvisited() {
for (int i = 0; i < sizeX; i++) {
for (int j = 0; j < sizeY; j++) {
if (!cells[i][j].isVisited()) {
return true;
}
}
}
return false;
}

public Cell backtrack() {
if (backtracker.size() == 0) return null;
Cell cell = backtracker.get(backtracker.size() - 1);
backtracker.remove(cell);
return cell;
}

public Cell getCell(int x, int y) {
return cells[x][y];
}

}

public class Cell {

private Maze maze;
private int x, y;
private boolean visited = false;
// top, right, bottom. left
private boolean[] walls = new boolean[]{true, true, true, true};

public Cell(Maze maze, int x, int y) {
this.maze = maze;
this.x = x;
this.y = y;
}

public Cell pickNext() {
List<Cell> neighbors = new ArrayList<>();
if (y != maze.sizeY - 1) neighbors.add(maze.getCell(x, y + 1));
else neighbors.add(null);
if (x != maze.sizeX - 1) neighbors.add(maze.getCell(x + 1, y));
else neighbors.add(null);
if (y != 0) neighbors.add(maze.getCell(x, y - 1));
else neighbors.add(null);
if (x != 0) neighbors.add(maze.getCell(x - 1, y));
else neighbors.add(null);
boolean hasUnvisitedNeighbor = false;
for (Cell c : neighbors) {
if (c == null) continue;
if (!c.isVisited()) hasUnvisitedNeighbor = true;
}
if (hasUnvisitedNeighbor) {
int random = (int) Math.floor(Math.random() * 4);
Cell next = neighbors.get(random);
while (next == null || next.isVisited()) {
random = (int) Math.floor(Math.random() * 4);
next = neighbors.get(random);
}
this.breakWall(random);
next.breakWall((random + 2) % 4);
return next;
} else return maze.backtrack();
}

public void breakWall(int wall) {
walls[wall] = false;
}

public void visit() {
visited = true;
}

public boolean isVisited() {
return visited;
}

}

关于java - 使用DFS回溯算法生成迷宫的问题,我们在Stack Overflow上找到一个类似的问题: https://stackoverflow.com/questions/60368575/

26 4 0
Copyright 2021 - 2024 cfsdn All Rights Reserved 蜀ICP备2022000587号
广告合作:1813099741@qq.com 6ren.com