我有一个nxm个由非负整数组成的矩阵。例如:

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

“投下炸弹”会使目标细胞及其所有八个邻居的数量减少一个,直到最小值为零。

x x x 
x X x
x x x

什么样的算法可以确定将所有细胞减少到零所需的最少炸弹数量?

B选项(因为我不是一个细心的读者)

事实上,问题的第一个版本并不是我要寻找的答案。我没有仔细阅读整个任务,有额外的约束条件,让我们说:

那么简单的问题是,当行中的序列必须是非递增的:

8 7 6 6 5是可能的输入序列

7 8 5 5 2是不可能的,因为7 -> 8在一个序列中增长。

也许为“简单”的问题找到答案会有助于为更难的问题找到解决方案。

PS:我相信当我们有几个相同的情况需要最少的炸弹来清除上面的线时,我们会选择在“左侧”使用最多炸弹的一个。还有什么证据是正确的吗?


当前回答

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))

其他回答

Pólya说:“如果你不能解决一个问题,那么有一个更容易解决的问题:找到它。”

显然更简单的问题是一维问题(当网格是单行时)。让我们从最简单的算法开始——贪婪地轰炸最大的目标。什么时候会出问题?

给定11 11,贪婪算法对先炸毁哪个单元格无关。当然,中心单元格更好——它一次将所有三个单元格归零。这就提出了一种新的算法a,“炸弹最小化剩余的总和”。这个算法什么时候会出错?

给定1 1 2 11 1,算法A在轰炸第2,第3或第4单元格之间是无所谓的。但是轰炸第二个单元格,留下0 0 11 11比轰炸第三个单元格,留下10 10 10 10 1好。如何解决这个问题?轰炸第三个单元格的问题是,左边的功和右边的功必须分开做。

“炸弹使剩余的总和最小化,但使左边(我们轰炸的地方)的最小值和右边的最小值最大化”如何?叫这个算法b,这个算法什么时候出错?


编辑:在阅读了评论之后,我同意一个更有趣的问题将是改变一维问题,使其两端连接起来。很乐意看到这方面的进展。

我想不出一个计算实际数字的方法除非用我最好的启发式方法计算轰炸行动并希望得到一个合理的结果。

So my method is to compute a bombing efficiency metric for each cell, bomb the cell with the highest value, .... iterate the process until I've flattened everything. Some have advocated using simple potential damage (i.e. score from 0 to 9) as a metric, but that falls short by pounding high value cells and not making use of damage overlap. I'd calculate cell value - sum of all neighbouring cells, reset any positive to 0 and use the absolute value of anything negative. Intuitively this metric should make a selection that help maximise damage overlap on cells with high counts instead of pounding those directly.

下面的代码在28个炸弹中达到了测试场的完全破坏(注意,使用潜在伤害作为度量,结果是31!)

using System;
using System.Collections.Generic;
using System.Linq;

namespace StackOverflow
{
  internal class Program
  {
    // store the battle field as flat array + dimensions
    private static int _width = 5;
    private static int _length = 7;
    private static int[] _field = new int[] {
        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
    };
    // this will store the devastation metric
    private static int[] _metric;

    // do the work
    private static void Main(string[] args)
    {
        int count = 0;

        while (_field.Sum() > 0)
        {
            Console.Out.WriteLine("Round {0}:", ++count);
            GetBlastPotential();
            int cell_to_bomb = FindBestBombingSite();
            PrintField(cell_to_bomb);
            Bomb(cell_to_bomb);
        }
        Console.Out.WriteLine("Done in {0} rounds", count);
    } 

    // convert 2D position to 1D index
    private static int Get1DCoord(int x, int y)
    {
        if ((x < 0) || (y < 0) || (x >= _width) || (y >= _length)) return -1;
        else
        {
            return (y * _width) + x;
        }
    }

    // Convert 1D index to 2D position
    private static void Get2DCoord(int n, out int x, out int y)
    {
        if ((n < 0) || (n >= _field.Length))
        {
            x = -1;
            y = -1;
        }
        else
        {
            x = n % _width;
            y = n / _width;
        }
    }

    // Compute a list of 1D indices for a cell neighbours
    private static List<int> GetNeighbours(int cell)
    {
        List<int> neighbours = new List<int>();
        int x, y;
        Get2DCoord(cell, out x, out y);
        if ((x >= 0) && (y >= 0))
        {
            List<int> tmp = new List<int>();
            tmp.Add(Get1DCoord(x - 1, y - 1));
            tmp.Add(Get1DCoord(x - 1, y));
            tmp.Add(Get1DCoord(x - 1, y + 1));
            tmp.Add(Get1DCoord(x, y - 1));
            tmp.Add(Get1DCoord(x, y + 1));
            tmp.Add(Get1DCoord(x + 1, y - 1));
            tmp.Add(Get1DCoord(x + 1, y));
            tmp.Add(Get1DCoord(x + 1, y + 1));

            // eliminate invalid coords - i.e. stuff past the edges
            foreach (int c in tmp) if (c >= 0) neighbours.Add(c);
        }
        return neighbours;
    }

    // Compute the devastation metric for each cell
    // Represent the Value of the cell minus the sum of all its neighbours
    private static void GetBlastPotential()
    {
        _metric = new int[_field.Length];
        for (int i = 0; i < _field.Length; i++)
        {
            _metric[i] = _field[i];
            List<int> neighbours = GetNeighbours(i);
            if (neighbours != null)
            {
                foreach (int j in neighbours) _metric[i] -= _field[j];
            }
        }
        for (int i = 0; i < _metric.Length; i++)
        {
            _metric[i] = (_metric[i] < 0) ? Math.Abs(_metric[i]) : 0;
        }
    }

    //// Compute the simple expected damage a bomb would score
    //private static void GetBlastPotential()
    //{
    //    _metric = new int[_field.Length];
    //    for (int i = 0; i < _field.Length; i++)
    //    {
    //        _metric[i] = (_field[i] > 0) ? 1 : 0;
    //        List<int> neighbours = GetNeighbours(i);
    //        if (neighbours != null)
    //        {
    //            foreach (int j in neighbours) _metric[i] += (_field[j] > 0) ? 1 : 0;
    //        }
    //    }            
    //}

    // Update the battle field upon dropping a bomb
    private static void Bomb(int cell)
    {
        List<int> neighbours = GetNeighbours(cell);
        foreach (int i in neighbours)
        {
            if (_field[i] > 0) _field[i]--;
        }
    }

    // Find the best bombing site - just return index of local maxima
    private static int FindBestBombingSite()
    {
        int max_idx = 0;
        int max_val = int.MinValue;
        for (int i = 0; i < _metric.Length; i++)
        {
            if (_metric[i] > max_val)
            {
                max_val = _metric[i];
                max_idx = i;
            }
        }
        return max_idx;
    }

    // Display the battle field on the console
    private static void PrintField(int cell)
    {
        for (int x = 0; x < _width; x++)
        {
            for (int y = 0; y < _length; y++)
            {
                int c = Get1DCoord(x, y);
                if (c == cell)
                    Console.Out.Write(string.Format("[{0}]", _field[c]).PadLeft(4));
                else
                    Console.Out.Write(string.Format(" {0} ", _field[c]).PadLeft(4));
            }
            Console.Out.Write(" || ");
            for (int y = 0; y < _length; y++)
            {
                int c = Get1DCoord(x, y);
                if (c == cell)
                    Console.Out.Write(string.Format("[{0}]", _metric[c]).PadLeft(4));
                else
                    Console.Out.Write(string.Format(" {0} ", _metric[c]).PadLeft(4));
            }
            Console.Out.WriteLine();
        }
        Console.Out.WriteLine();
    }           
  }
}

产生的轰炸模式输出如下(左边是字段值,右边是度量值)

Round 1:
  2   1   4   2   3   2   6  ||   7  16   8  10   4  18   6
  3   5   3   1   1   1   9  ||  11  18  18  21  17  28   5
  4  [2]  4   2   3   4   1  ||  19 [32] 21  20  17  24  22
  7   6   2   4   4   3   6  ||   8  17  20  14  16  22   8
  1   2   1   1   1   2   4  ||  14  15  14  11  13  16   7

Round 2:
  2   1   4   2   3   2   6  ||   5  13   6   9   4  18   6
  2   4   2   1   1  [1]  9  ||  10  15  17  19  17 [28]  5
  3   2   3   2   3   4   1  ||  16  24  18  17  17  24  22
  6   5   1   4   4   3   6  ||   7  14  19  12  16  22   8
  1   2   1   1   1   2   4  ||  12  12  12  10  13  16   7

Round 3:
  2   1   4   2   2   1   5  ||   5  13   6   7   3  15   5
  2   4   2   1   0   1   8  ||  10  15  17  16  14  20   2
  3  [2]  3   2   2   3   0  ||  16 [24] 18  15  16  21  21
  6   5   1   4   4   3   6  ||   7  14  19  11  14  19   6
  1   2   1   1   1   2   4  ||  12  12  12  10  13  16   7

Round 4:
  2   1   4   2   2   1   5  ||   3  10   4   6   3  15   5
  1   3   1   1   0   1   8  ||   9  12  16  14  14  20   2
  2   2   2   2   2  [3]  0  ||  13  16  15  12  16 [21] 21
  5   4   0   4   4   3   6  ||   6  11  18   9  14  19   6
  1   2   1   1   1   2   4  ||  10   9  10   9  13  16   7

Round 5:
  2   1   4   2   2   1   5  ||   3  10   4   6   2  13   3
  1   3   1   1   0  [0]  7  ||   9  12  16  13  12 [19]  2
  2   2   2   2   1   3   0  ||  13  16  15  10  14  15  17
  5   4   0   4   3   2   5  ||   6  11  18   7  13  17   6
  1   2   1   1   1   2   4  ||  10   9  10   8  11  13   5

Round 6:
  2   1   4   2   1   0   4  ||   3  10   4   5   2  11   2
  1   3   1   1   0   0   6  ||   9  12  16  11   8  13   0
  2   2   2   2   0   2   0  ||  13  16  15   9  14  14  15
  5   4  [0]  4   3   2   5  ||   6  11 [18]  6  11  15   5
  1   2   1   1   1   2   4  ||  10   9  10   8  11  13   5

Round 7:
  2   1   4   2   1   0   4  ||   3  10   4   5   2  11   2
  1   3   1   1   0   0   6  ||   8  10  13   9   7  13   0
  2  [1]  1   1   0   2   0  ||  11 [15] 12   8  12  14  15
  5   3   0   3   3   2   5  ||   3   8  10   3   8  15   5
  1   1   0   0   1   2   4  ||   8   8   7   7   9  13   5

Round 8:
  2   1   4   2   1   0   4  ||   1   7   2   4   2  11   2
  0   2   0   1   0   0   6  ||   7   7  12   7   7  13   0
  1   1   0   1   0   2   0  ||   8   8  10   6  12  14  15
  4   2   0   3   3  [2]  5  ||   2   6   8   2   8 [15]  5
  1   1   0   0   1   2   4  ||   6   6   6   7   9  13   5

Round 9:
  2   1   4   2   1   0   4  ||   1   7   2   4   2  11   2
  0   2   0   1   0   0   6  ||   7   7  12   7   6  12   0
  1   1   0   1   0  [1]  0  ||   8   8  10   5  10 [13] 13
  4   2   0   3   2   2   4  ||   2   6   8   0   6   9   3
  1   1   0   0   0   1   3  ||   6   6   6   5   8  10   4

Round 10:
  2   1   4   2   1   0   4  ||   1   7   2   4   2  10   1
  0   2  [0]  1   0   0   5  ||   7   7 [12]  7   6  11   0
  1   1   0   1   0   1   0  ||   8   8  10   4   8   9  10
  4   2   0   3   1   1   3  ||   2   6   8   0   6   8   3
  1   1   0   0   0   1   3  ||   6   6   6   4   6   7   2

Round 11:
  2   0   3   1   1   0   4  ||   0   6   0   3   0  10   1
  0   1   0   0   0  [0]  5  ||   4   5   5   5   3 [11]  0
  1   0   0   0   0   1   0  ||   6   8   6   4   6   9  10
  4   2   0   3   1   1   3  ||   1   5   6   0   5   8   3
  1   1   0   0   0   1   3  ||   6   6   6   4   6   7   2

Round 12:
  2   0   3   1   0   0   3  ||   0   6   0   2   1   7   1
  0   1   0   0   0   0   4  ||   4   5   5   4   1   7   0
  1   0   0   0   0  [0]  0  ||   6   8   6   4   5  [9]  8
  4   2   0   3   1   1   3  ||   1   5   6   0   4   7   2
  1   1   0   0   0   1   3  ||   6   6   6   4   6   7   2

Round 13:
  2   0   3   1   0   0   3  ||   0   6   0   2   1   6   0
  0   1   0   0   0   0   3  ||   4   5   5   4   1   6   0
  1  [0]  0   0   0   0   0  ||   6  [8]  6   3   3   5   5
  4   2   0   3   0   0   2  ||   1   5   6   0   4   6   2
  1   1   0   0   0   1   3  ||   6   6   6   3   4   4   0

Round 14:
  2   0   3   1   0  [0]  3  ||   0   5   0   2   1  [6]  0
  0   0   0   0   0   0   3  ||   2   5   4   4   1   6   0
  0   0   0   0   0   0   0  ||   4   4   4   3   3   5   5
  3   1   0   3   0   0   2  ||   0   4   5   0   4   6   2
  1   1   0   0   0   1   3  ||   4   4   5   3   4   4   0

Round 15:
  2   0   3   1   0   0   2  ||   0   5   0   2   1   4   0
  0   0   0   0   0   0   2  ||   2   5   4   4   1   4   0
  0   0   0   0   0   0   0  ||   4   4   4   3   3   4   4
  3   1   0   3   0  [0]  2  ||   0   4   5   0   4  [6]  2
  1   1   0   0   0   1   3  ||   4   4   5   3   4   4   0

Round 16:
  2  [0]  3   1   0   0   2  ||   0  [5]  0   2   1   4   0
  0   0   0   0   0   0   2  ||   2   5   4   4   1   4   0
  0   0   0   0   0   0   0  ||   4   4   4   3   3   3   3
  3   1   0   3   0   0   1  ||   0   4   5   0   3   3   1
  1   1   0   0   0   0   2  ||   4   4   5   3   3   3   0

Round 17:
  1   0   2   1   0   0   2  ||   0   3   0   1   1   4   0
  0   0   0   0   0   0   2  ||   1   3   3   3   1   4   0
  0   0   0   0   0   0   0  ||   4   4   4   3   3   3   3
  3   1  [0]  3   0   0   1  ||   0   4  [5]  0   3   3   1
  1   1   0   0   0   0   2  ||   4   4   5   3   3   3   0

Round 18:
  1   0   2   1   0   0   2  ||   0   3   0   1   1   4   0
  0   0   0   0   0   0   2  ||   1   3   3   3   1   4   0
  0   0   0   0   0   0   0  ||   3   3   2   2   2   3   3
  3  [0]  0   2   0   0   1  ||   0  [4]  2   0   2   3   1
  1   0   0   0   0   0   2  ||   2   4   2   2   2   3   0

Round 19:
  1   0   2   1   0  [0]  2  ||   0   3   0   1   1  [4]  0
  0   0   0   0   0   0   2  ||   1   3   3   3   1   4   0
  0   0   0   0   0   0   0  ||   2   2   2   2   2   3   3
  2   0   0   2   0   0   1  ||   0   2   2   0   2   3   1
  0   0   0   0   0   0   2  ||   2   2   2   2   2   3   0

Round 20:
  1  [0]  2   1   0   0   1  ||   0  [3]  0   1   1   2   0
  0   0   0   0   0   0   1  ||   1   3   3   3   1   2   0
  0   0   0   0   0   0   0  ||   2   2   2   2   2   2   2
  2   0   0   2   0   0   1  ||   0   2   2   0   2   3   1
  0   0   0   0   0   0   2  ||   2   2   2   2   2   3   0

Round 21:
  0   0   1   1   0   0   1  ||   0   1   0   0   1   2   0
  0   0   0   0   0   0   1  ||   0   1   2   2   1   2   0
  0   0   0   0   0   0   0  ||   2   2   2   2   2   2   2
  2   0   0   2   0  [0]  1  ||   0   2   2   0   2  [3]  1
  0   0   0   0   0   0   2  ||   2   2   2   2   2   3   0

Round 22:
  0   0   1   1   0   0   1  ||   0   1   0   0   1   2   0
  0   0   0   0   0   0   1  ||   0   1   2   2   1   2   0
 [0]  0   0   0   0   0   0  ||  [2]  2   2   2   2   1   1
  2   0   0   2   0   0   0  ||   0   2   2   0   2   1   1
  0   0   0   0   0   0   1  ||   2   2   2   2   2   1   0

Round 23:
  0   0   1   1   0   0   1  ||   0   1   0   0   1   2   0
  0   0  [0]  0   0   0   1  ||   0   1  [2]  2   1   2   0
  0   0   0   0   0   0   0  ||   1   1   2   2   2   1   1
  1   0   0   2   0   0   0  ||   0   1   2   0   2   1   1
  0   0   0   0   0   0   1  ||   1   1   2   2   2   1   0

Round 24:
  0   0   0   0   0   0   1  ||   0   0   0   0   0   2   0
  0   0   0   0   0   0   1  ||   0   0   0   0   0   2   0
  0   0  [0]  0   0   0   0  ||   1   1  [2]  2   2   1   1
  1   0   0   2   0   0   0  ||   0   1   2   0   2   1   1
  0   0   0   0   0   0   1  ||   1   1   2   2   2   1   0

Round 25:
  0   0   0   0   0  [0]  1  ||   0   0   0   0   0  [2]  0
  0   0   0   0   0   0   1  ||   0   0   0   0   0   2   0
  0   0   0   0   0   0   0  ||   1   1   1   1   1   1   1
  1   0   0   1   0   0   0  ||   0   1   1   0   1   1   1
  0   0   0   0   0   0   1  ||   1   1   1   1   1   1   0

Round 26:
  0   0   0   0   0   0   0  ||   0   0   0   0   0   0   0
  0   0   0   0   0   0   0  ||   0   0   0   0   0   0   0
 [0]  0   0   0   0   0   0  ||  [1]  1   1   1   1   0   0
  1   0   0   1   0   0   0  ||   0   1   1   0   1   1   1
  0   0   0   0   0   0   1  ||   1   1   1   1   1   1   0

Round 27:
  0   0   0   0   0   0   0  ||   0   0   0   0   0   0   0
  0   0   0   0   0   0   0  ||   0   0   0   0   0   0   0
  0   0  [0]  0   0   0   0  ||   0   0  [1]  1   1   0   0
  0   0   0   1   0   0   0  ||   0   0   1   0   1   1   1
  0   0   0   0   0   0   1  ||   0   0   1   1   1   1   0

Round 28:
  0   0   0   0   0   0   0  ||   0   0   0   0   0   0   0
  0   0   0   0   0   0   0  ||   0   0   0   0   0   0   0
  0   0   0   0   0   0   0  ||   0   0   0   0   0   0   0
  0   0   0   0   0  [0]  0  ||   0   0   0   0   0  [1]  1
  0   0   0   0   0   0   1  ||   0   0   0   0   0   1   0

Done in 28 rounds

这里有一个解决方案,推广良好的性质的角。

让我们假设我们可以为给定的字段找到一个完美的落点,也就是说,一个减少其中值的最佳方法。然后,为了找到最少的炸弹数量,一个算法的初稿可能是(代码是从ruby实现中复制粘贴的):

dropped_bomb_count = 0
while there_are_cells_with_non_zero_count_left
  coordinates = choose_a_perfect_drop_point
  drop_bomb(coordinates)
  dropped_bomb_count += 1
end
return dropped_bomb_count

挑战是choose_a_perfect_drop_point。首先,让我们定义一个完美的落点是什么。

(x, y)的放置点会减少(x, y)中的值。它也可能会减少其他单元格中的值。 (x, y)的放置点A比(x, y)的放置点b更好,如果它减少了b所减少的单元格的适当超集中的值。 如果没有其他更好的投放点,投放点是最大的。 (x, y)的两个放置点是等效的,如果它们减少了同一组单元格。 如果(x, y)的放置点等价于(x, y)的所有最大放置点,那么它就是完美的。

如果(x, y)存在一个完美的投放点,那么您不能比在(x, y)的一个完美投放点上投放炸弹更有效地降低(x, y)处的值。

给定字段的完美放置点是其任何单元格的完美放置点。

以下是一些例子:

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

单元格(0,0)(从零开始的索引)的完美放置点是(1,1)。(1,1)的所有其他放置点,即(0,0)、(0,1)和(1,0),减少的单元格较少。

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

单元格(2,2)(从零开始的索引)的完美落点是(2,2),以及所有周围的单元格(1,1)、(1,2)、(1,3)、(2,1)、(2,3)、(3,1)、(3,2)和(3,3)。

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

单元格(2,2)的完美放置点是(3,1):它减少了(2,2)中的值和(4,0)中的值。(2,2)的所有其他放置点都不是最大的,因为它们减少了一个单元格。(2,2)的完美下拉点也是(4,0)的完美下拉点,它是字段的唯一完美下拉点。它为这个领域带来了完美的解决方案(一颗炸弹)。

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

(2,2)没有完美的落点:(1,1)和(1,3)都减少(2,2)和另一个单元格(它们是(2,2)的最大落点),但它们不相等。然而,(1,1)是(0,0)的完美落点,(1,3)是(0,4)的完美落点。

根据完美落点的定义和一定的检查顺序,我得到了以下问题示例的结果:

Drop bomb on 1, 1
Drop bomb on 1, 1
Drop bomb on 1, 5
Drop bomb on 1, 5
Drop bomb on 1, 5
Drop bomb on 1, 6
Drop bomb on 1, 2
Drop bomb on 1, 2
Drop bomb on 0, 6
Drop bomb on 0, 6
Drop bomb on 2, 1
Drop bomb on 2, 5
Drop bomb on 2, 5
Drop bomb on 2, 5
Drop bomb on 3, 1
Drop bomb on 3, 0
Drop bomb on 3, 0
Drop bomb on 3, 0
Drop bomb on 3, 0
Drop bomb on 3, 0
Drop bomb on 3, 4
Drop bomb on 3, 4
Drop bomb on 3, 3
Drop bomb on 3, 3
Drop bomb on 3, 6
Drop bomb on 3, 6
Drop bomb on 3, 6
Drop bomb on 4, 6
28

然而,该算法只有在每一步之后至少有一个完美落点时才能工作。可以在没有完美落点的情况下构建例子:

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

对于这些情况,我们可以修改算法,这样我们就不会选择完美的落点,而是选择一个具有最大落点的最小选择的坐标,然后计算每个选择的最小值。在上面的例子中,所有有值的单元格都有两个最大落点。例如,(0,1)有最大落点(1,1)和(1,2)。选择其中任何一个,然后计算最小值,会得到这样的结果:

Drop bomb on 1, 1
Drop bomb on 2, 2
Drop bomb on 1, 2
Drop bomb on 2, 1
2

对于更新后的问题,简单的贪心算法可以得到最优结果。

向单元格A[1,1]投掷A[0,0]炸弹,然后向单元格A[2,1]投掷A[1,0]炸弹,并继续向下此过程。要清除左下角,向单元格A[n -2,1]投掷max(A[n -1,0], A[n -2,0], A[n -3,0])炸弹。这将完全清除前3列。

用同样的方法清除第3、4、5列,然后是第6、7、8列,等等。

不幸的是,这并不能帮助找到最初问题的解决方案。


“更大”的问题(没有“非增加”约束)可能被证明是np困难的。这是证明的草图。

假设我们有一个度为3的平面图形。我们来求这个图的最小顶点覆盖。根据维基百科的文章,这个问题对于3次以下的平面图形是np困难的。这可以通过平面3SAT的简化来证明。平面3SAT的硬度由3SAT降低而成。这两个证明都在Erik Demaine教授最近的“算法下界”讲座(第7和第9讲)中提出。

如果我们分割原始图的一些边(图中左边的图),每条边都有偶数个额外的节点,结果图(图中右边的图)应该对原始顶点具有完全相同的最小顶点覆盖。这样的转换允许将图顶点对齐到网格上的任意位置。

如果我们将图顶点只放置在偶数行和列上(这样就不会有两条边与一个顶点形成锐角),在有边的地方插入“1”,在其他网格位置插入“0”,我们可以使用原始问题的任何解决方案来找到最小顶点覆盖。

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.