这可以用深度为O(3^(n))的树来求解。其中n是所有平方和。

首先考虑用O(9^n)树来解决问题是很简单的，只需考虑所有可能的爆炸位置。有关示例，请参阅Alfe的实现。

接下来我们意识到，我们可以从下往上轰炸，仍然得到一个最小的轰炸模式。

Start from the bottom left corner. Bomb it to oblivion with the only plays that make sense (up and to the right). Move one square to the right. While the target has a value greater than zero, consider each of the 2 plays that make sense (straight up or up and to the right), reduce the value of the target by one, and make a new branch for each possibility. Move another to the right. While the target has a value greater than zero, consider each of the 3 plays that make sense (up left, up, and up right), reduce the value of the target by one, and make a new branch for each possibility. Repeat steps 5 and 6 until the row is eliminated. Move up a row and repeat steps 1 to 7 until the puzzle is solved.

这个算法是正确的，因为

有必要在某一时刻完成每一行。 完成一行总是需要一个游戏，一个在上面，一个在下面，或者在这一行内。 选择在未清除的最低行之上的玩法总是比选择在该行之上或该行之下的玩法更好。

在实践中，这个算法通常会比它的理论最大值做得更好，因为它会定期轰炸邻居并减少搜索的大小。如果我们假设每次轰炸都会减少4个额外目标的价值，那么我们的算法将运行在O(3^(n/4))或大约O(1.3^n)。

Because this algorithm is still exponential, it would be wise to limit the depth of the search. We might limit the number of branches allowed to some number, X, and once we are this deep we force the algorithm to choose the best path it has identified so far (the one that has the minimum total board sum in one of its terminal leaves). Then our algorithm is guaranteed to run in O(3^X) time, but it is not guaranteed to get the correct answer. However, we can always increase X and test empirically if the trade off between increased computation and better answers is worthwhile.

这是一个广度搜索，通过这个“迷宫”的位置寻找最短路径(一系列轰炸)。不，我不能证明没有更快的算法，抱歉。

#!/usr/bin/env python

M = ((1,2,3,4),
     (2,3,4,5),
     (5,2,7,4),
     (2,3,5,8))

def eachPossibleMove(m):
  for y in range(1, len(m)-1):
    for x in range(1, len(m[0])-1):
      if (0 == m[y-1][x-1] == m[y-1][x] == m[y-1][x+1] ==
               m[y][x-1]   == m[y][x]   == m[y][x+1] ==
               m[y+1][x-1] == m[y+1][x] == m[y+1][x+1]):
        continue
      yield x, y

def bomb(m, (mx, my)):
  return tuple(tuple(max(0, m[y][x]-1)
      if mx-1 <= x <= mx+1 and my-1 <= y <= my+1
      else m[y][x]
      for x in range(len(m[y])))
    for y in range(len(m)))

def findFirstSolution(m, path=[]):
#  print path
#  print m
  if sum(map(sum, m)) == 0:  # empty?
    return path
  for move in eachPossibleMove(m):
    return findFirstSolution(bomb(m, move), path + [ move ])

def findShortestSolution(m):
  black = {}
  nextWhite = { m: [] }
  while nextWhite:
    white = nextWhite
    nextWhite = {}
    for position, path in white.iteritems():
      for move in eachPossibleMove(position):
        nextPosition = bomb(position, move)
        nextPath = path + [ move ]
        if sum(map(sum, nextPosition)) == 0:  # empty?
          return nextPath
        if nextPosition in black or nextPosition in white:
          continue  # ignore, found that one before
        nextWhite[nextPosition] = nextPath

def main(argv):
  if argv[1] == 'first':
    print findFirstSolution(M)
  elif argv[1] == 'shortest':
    print findShortestSolution(M)
  else:
    raise NotImplementedError(argv[1])

if __name__ == '__main__':
  import sys
  sys.exit(main(sys.argv))

使用分支和定界的数学整数线性规划

As it has already been mentioned, this problem can be solved using integer linear programming (which is NP-Hard). Mathematica already has ILP built in. "To solve an integer linear programming problem Mathematica first solves the equational constraints, reducing the problem to one containing inequality constraints only. Then it uses lattice reduction techniques to put the inequality system in a simpler form. Finally, it solves the simplified optimization problem using a branch-and-bound method." [see Constrained Optimization Tutorial in Mathematica.. ]

我写了下面的代码，利用ILP库的Mathematica。它的速度快得惊人。

solveMatrixBombProblem[problem_, r_, c_] := 
 Module[{}, 
  bombEffect[x_, y_, m_, n_] := 
   Table[If[(i == x || i == x - 1 || i == x + 1) && (j == y || 
        j == y - 1 || j == y + 1), 1, 0], {i, 1, m}, {j, 1, n}];
  bombMatrix[m_, n_] := 
   Transpose[
    Table[Table[
      Part[bombEffect[(i - Mod[i, n])/n + 1, Mod[i, n] + 1, m, 
        n], (j - Mod[j, n])/n + 1, Mod[j, n] + 1], {j, 0, 
       m*n - 1}], {i, 0, m*n - 1}]];
  X := x /@ Range[c*r];
  sol = Minimize[{Total[X], 
     And @@ Thread[bombMatrix[r, c].X >= problem] && 
      And @@ Thread[X >= 0] && Total[X] <= 10^100 && 
      Element[X, Integers]}, X];
  Print["Minimum required bombs = ", sol[[1]]];
  Print["A possible solution = ", 
   MatrixForm[
    Table[x[c*i + j + 1] /. sol[[2]], {i, 0, r - 1}, {j, 0, 
      c - 1}]]];]

对于问题中提供的示例:

solveMatrixBombProblem[{2, 3, 4, 7, 1, 1, 5, 2, 6, 2, 4, 3, 4, 2, 1, 2, 1, 2, 4, 1, 3, 1, 3, 4, 1, 2, 1, 4, 3, 2, 6, 9, 1, 6, 4}, 7, 5]

输出

对于那些用贪婪算法读这篇文章的人

在下面这个10x10的问题上试试你的代码:

5   20  7   1   9   8   19  16  11  3  
17  8   15  17  12  4   5   16  8   18  
4   19  12  11  9   7   4   15  14  6  
17  20  4   9   19  8   17  2   10  8  
3   9   10  13  8   9   12  12  6   18  
16  16  2   10  7   12  17  11  4   15  
11  1   15  1   5   11  3   12  8   3  
7   11  16  19  17  11  20  2   5   19  
5   18  2   17  7   14  19  11  1   6  
13  20  8   4   15  10  19  5   11  12

这里用逗号分隔:

5, 20, 7, 1, 9, 8, 19, 16, 11, 3, 17, 8, 15, 17, 12, 4, 5, 16, 8, 18, 4, 19, 12, 11, 9, 7, 4, 15, 14, 6, 17, 20, 4, 9, 19, 8, 17, 2, 10, 8, 3, 9, 10, 13, 8, 9, 12, 12, 6, 18, 16, 16, 2, 10, 7, 12, 17, 11, 4, 15, 11, 1, 15, 1, 5, 11, 3, 12, 8, 3, 7, 11, 16, 19, 17, 11, 20, 2, 5, 19, 5, 18, 2, 17, 7, 14, 19, 11, 1, 6, 13, 20, 8, 4, 15, 10, 19, 5, 11, 12

对于这个问题，我的解决方案包含208个炸弹。这里有一个可能的解决方案(我能够在大约12秒内解决这个问题)。

作为一种测试Mathematica产生结果的方法，看看你的贪婪算法是否能做得更好。

这将是一个贪婪的方法:

计算一个阶为n X m的“score”矩阵，其中score[i][j]是如果位置(i,j)被炸毁，则矩阵中各点的总扣除额。(一个点的最高分数是9分，最低分数是0分) 逐行移动，找到并选择第一个得分最高的位置(例如(i,j))。 炸弹(i, j)。增加炸弹数量。 如果原矩阵的所有元素都不为零，则转到1。

但我怀疑这是否是最佳解决方案。

编辑:

我上面提到的贪心方法，虽然有效，但很可能不能给我们最优的解决方案。所以我想应该添加一些DP的元素。

我想我们可以同意，在任何时候，具有最高“分数”(分数[I][j] =总扣分，如果(I,j)被炸)的位置之一必须被瞄准。从这个假设开始，下面是新的方法:

NumOfBombs(M):(返回所需的最小炸弹数量)

给定一个矩阵M (n X M)，如果M中的所有元素都为0，则返回0。 计算“分数”矩阵M。 设k个不同的位置P1 P2…Pk (1 <= k <= n*m)，为m中得分最高的位置。 return (1 + min(NumOfBombs(M1)， NumOfBombs(M2)，…， NumOfBombs(Mk)) M1, M2,……，Mk是我们轰炸位置P1, P2，…， Pk。

此外，如果我们想在此基础上破坏位置的顺序，我们必须跟踪“min”的结果。

如果你想要绝对最优解来清理棋盘，你将不得不使用经典的回溯，但如果矩阵非常大，它将需要很长时间才能找到最佳解，如果你想要一个“可能的”最优解，你可以使用贪婪算法，如果你需要帮助写算法，我可以帮助你

现在想想，这是最好的办法。在那里制作另一个矩阵，存储通过投掷炸弹而移除的点，然后选择点数最多的单元格，并在那里投掷炸弹更新点数矩阵，然后继续。例子:

2 3 5 -> (2+(1*3)) (3+(1*5)) (5+(1*3))
1 3 2 -> (1+(1*4)) (3+(1*7)) (2+(1*4))
1 0 2 -> (1+(1*2)) (0+(1*5)) (2+(1*2))

对于每个相邻的高于0的单元格，单元格值+1

在这里，线性规划方法似乎非常有用。

设Pm x n为包含位置值的矩阵:

现在定义一个炸弹矩阵B(x, y)m x n，其中1≤x≤m, 1≤y≤n如下所示

以这样一种方式

例如:

所以我们正在寻找一个矩阵Bm x n = [bij]

可以定义为炸弹矩阵的和: (qij将是我们在pij位置投放的炸弹数量) pij - bij≤0(为了更简洁，我们称之为P - B≤0)

而且，B应该使和最小。

我们也可以把B写成前面的丑矩阵:

由于P - B≤0(即P≤B)，我们得到了如下线性不等式系统:

qmn x1定义为

PMN x 1定义为

我们可以说我们有一个方程组是smnxmn这个矩阵要倒转来解方程组。我自己没有扩展它，但我相信在代码中应该很容易做到。

现在，我们有一个最小的问题可以表述为

I believe it is something easy, almost trivial to be solved with something like the simplex algorithm (there is this rather cool doc about it). However, I do know almost no linear programming (I will take a course about it on Coursera but it is just in the future...), I had some headaches trying to understand it and I have a huge freelance job to finish so I just give up here. It can be that I did something wrong at some point, or that it can't go any further, but I believe this path can eventually lead to the solution. Anyway, I am anxious for your feedback.

(特别感谢这个神奇的网站从LaTeX表达式创建图片)