挑战：

给定0和1的矩阵（或2d数组），输出Conway的生活游戏达到稳定状态所需的步数，如果从未达到1，则输出-1。稳定状态是每个步骤都没有打开或关闭任何单元的状态。游戏必须在给定的矩阵中运行，顶部和底部连接，侧面连接。（即给定的4x3矩阵，它应该在4x3的圆环上运行）输入矩阵的大小不得超过15x15。

注意：如果矩阵以稳定状态开始，则输出应为0。

样品：

输入：

[[0,0,0],  
 [0,1,1],  
 [0,1,0]]

输出：

2

流程：（此不需要显示）

[[0,0,0],
 [0,1,1],
 [0,1,0]]

[[1,1,1],
 [1,1,1],
 [1,1,1]]

[[0,0,0],
 [0,0,0],
 [0,0,0]]

输入：

[[0,0,1,1],
 [0,1,1,1],
 [0,1,0,0],
 [0,1,1,1]]

输出：

2

处理：

[[0,0,1,1],
 [0,1,1,1],
 [0,1,0,0],
 [0,1,1,1]]

[[0,0,0,0],
 [0,1,0,1],
 [0,0,0,0],
 [0,1,0,1]]

[[0,0,0,0],
 [0,0,0,0],
 [0,0,0,0],
 [0,0,0,0]]

输入：

[[0,1,0,0],
 [0,1,0,0],
 [0,1,0,0],
 [0,0,0,0]]

输出：

-1

处理：

[[0,1,0,0],
 [0,1,0,0],
 [0,1,0,0],
 [0,0,0,0]]

[[0,0,0,0],
 [1,1,1,0],
 [0,0,0,0],
 [0,0,0,0]]

[[0,1,0,0],
 [0,1,0,0],
 [0,1,0,0],
 [0,0,0,0]]

永远重复

输入：

[[0,0,0,0],
 [0,0,0,1],
 [0,1,1,1],
 [0,0,1,0]]

输出：

4

处理：

[[0,0,0,0],
 [0,0,0,1],
 [0,1,1,1],
 [0,0,1,0]]

[[0,0,0,0],
 [1,0,0,1],
 [1,1,0,1],
 [0,1,1,1]]

[[0,1,0,0],
 [0,1,1,1],
 [0,0,0,0],
 [0,1,0,1]]

[[0,1,0,1],
 [1,1,1,0],
 [0,1,0,1],
 [1,0,1,0]]

[[0,0,0,0],
 [0,0,0,0],
 [0,0,0,0],
 [0,0,0,0]]

输入：

[[0,0,0,0],
 [0,1,1,0],
 [0,1,1,0],
 [0,0,0,0]]

输出：

0

处理：

开始状态是稳定的。

生命游戏规则

如果关闭（0）的一个单元正好紧靠三个（1）单元，则将其打开。否则，它将保留。如果打开的单元格在2或3个正方形的旁边，则表示打开。否则，它将关闭。

数学，130个 129字节

#&@@FirstPosition[Partition[CellularAutomaton[{224,{2,{{2,2,2},{2,1,2},{2,2,2}}},{1,1}},#,2^Length[Join@@#]],2,1],{x_,x_},0,1]-1&

我不建议尝试超过4x4的输入，因为这将永远花费很多时间（而且很多时间内存）。

说明

这只是简单地模拟了2 ^N步的人生游戏，其中N是输入中的单元数。这样可以保证，如果系统稳定下来，我们就可以达到稳定状态。然后，我们在模拟历史记录中找到第一对连续的相同状态。

让我们看一下代码：

2^Length[Join@@#]

由于用于平铺2D列表，因此计算2 ^N。Join@@

CellularAutomaton[{224,{2,{{2,2,2},{2,1,2},{2,2,2}}},{1,1}},#,...]

这模拟了《生命游戏2》 ^N代。3x3矩阵指定了一个整体二维自动机的邻域，并且224是标准“生命游戏”的规则号。我已经在Mathematica.SE上写了关于如何计算此数字的文章。

Partition[...,2,1]

这将获得所有连续（重叠）的世代对。

FirstPosition[...,{x_,x_},0,1]

查找第一对相同的世代，默认为 0为找不到，并将搜索范围限制为depth 1。如果这样一对被发现，结果以列表返回虽然。因此我们使用：

#&@@...

要从该列表中提取第一个元素（的默认值0是atomic，不受此影响）。

...-1

最后，我们减去一个，因为挑战期望0基于索引的索引和-1失败索引。

Lua中，531 509 488 487 464 424 405 404个字节

谁想要大量提交？\ o /

编辑：改进，但不知道如何打高尔夫球了，所以... 解释即将添加注释:)

在@ KennyLau的帮助下保存了约60个字节

小高尔夫球切削多一个字节通过重命名a成Y，以防止内联进制转换

function f(m)c={}t=table.concat::z::c[#c+1]={}p=c[#c]Y={}for i=1,#m do k=m[i]p[#p+1]=t(k)Y[i]={}for j=1,#k
do v=m[i%#m+1]l=j%#k+1w=m[(i-2)%#m+1]h=(j-2)%#k+1Y[i][j]=v[l]+k[l]+w[l]+v[h]+v[j]+w[h]+w[j]+k[h]end
end s=''for i=1,#m do k=m[i]for j=1,k do
x=Y[i][j]k[j]=k[j]>0 and((x<2or x>3)and 0or 1)or (x==3 and 1or 0)end
s=s..t(k)end for i=1,#c do
if(s==t(c[i]))then return#c>i and-1or i-1
end end goto z end

不打高尔夫球

function f(m)                -- takes a 2D array of 0 and 1s as input
  c={}                       -- intialise c -> contains a copy of each generation
  t=table.concat             -- shorthand for the concatenating function 
  ::z::                      -- label z, used to do an infinite loop
    c[#c+1]={}               -- initialise the first copy 
    p=c[#c]                  -- initialise a pointer to this copy
    a={}                     -- initialise the 2D array of adjacency
    for i=1,#m               -- iterate over the lines of m
    do
      k=m[i]                 -- shorthand for the current line
      p[#p+1]=t(k])          -- saves the current line of m as a string
      a[i]={}                -- initialise the array of adjacency for the current line
      for j=1,#k             -- iterate over each row of m
      do
                             -- the following statements are used to wraps at borders
        v=m[i%#m+1]          -- wrap bottom to top
        l=j%#k+1             -- wrap right to left
        w=m[(i-2)%#m+1]      -- wrap top to bottom
        h=(j-2)%#k+1         -- wrap left to right

        a[i][j]= v[l]        -- living cells are 1 and deads are 0
                +k[l]        -- just add the values of adjacent cells
                +w[l]        -- to know the number of alive adjacent cells
                +v[h]
                +v[j]
                +w[h]
                +w[j]
                +k[h]
      end
    end

    s=''                     -- s will be the representation of the current generation
    for i=1,#m               -- iterate over each line
    do
      k=m[i]                 -- shorthand for the current line
      for j=1,#k             -- iterate over each row
      do
        x=a[i][j]            -- shorthand for the number of adjacent to the current cell
                             -- the next line change the state of the current cell
        k[j]=k[j]>0          -- if it is alive
                and((x<2     --   and it has less than 2 adjacent
                    or x>3)  --   or more than 3 adjacent
                  and 0      --     kill it
                  or 1)      --     else let it alive
                or           -- if it is dead
                  (x==3      --   and it has 3 adjacent
                  and 1      --     give life to it
                  or 0)      --     else let it dead
      end
      s=s..t(k)              -- save the representation of the current line
    end
    for i=1,#c               -- iterate over all the generation done until now
    do                       
      if(s==t(c[i]))         -- if the representation of the current generation
      then                   -- is equal to one we saved
        return#c>i           -- check if it is the latest generation
              and-1          -- if it isn't, it means we are in a loop -> return -1
              or i-1         -- if it is, we did 2 generations without changing
                             --  -> return the number of generation
      end
    end
  goto z                     -- if we reach that point, loop back to the label z
end

测试用例

这是一些测试案例

function f(m)c={}t=table.concat::z::c[#c+1]={}p=c[#c]a={}for i=1,#m do k=m[i]p[#p+1]=t(k)a[i]={}for j=1,#k
do v=m[i%#m+1]l=j%#k+1w=m[(i-2)%#m+1]h=(j-2)%#k+1
a[i][j]=v[l]+k[l]+w[l]+v[h]+v[j]+w[h]+w[j]+k[h]end
end s=''for i=1,#m do k=m[i]for j=1,k do
x=a[i][j]k[j]=k[j]>0 and((x<2or x>3)and 0or 1)or (x==3 and 1or 0)end
s=s..t(k)end for i=1,#c do
if(s==t(c[i]))then return#c>i and-1or i-1
end end goto z end




print(f({{0,0,0},{0,1,1},{0,1,0}}))
print(f({{0,1,0,0},{0,1,0,0},{0,1,0,0},{0,0,0,0}}))
-- 53 generation, 15x15, takes 50-100 ms on a bad laptop
print(f({{0,0,0,0,1,1,0,1,0,0,0,0,1,0,0},
       {0,1,1,0,1,1,1,1,1,1,1,0,0,0,0},
       {0,1,1,1,0,1,0,1,0,0,0,0,1,0,0},
       {0,0,1,0,1,1,1,0,0,1,1,1,0,1,1},
       {1,1,0,0,1,1,1,0,1,1,0,0,0,1,0},
       {0,0,0,0,1,1,0,1,0,0,0,0,1,0,0},
       {0,1,1,0,1,1,1,1,1,1,1,0,0,0,0},
       {0,1,1,1,0,1,0,1,0,0,0,0,1,0,0},
       {0,0,1,0,1,1,1,0,0,1,1,1,0,1,1},
       {1,1,0,0,1,1,1,0,1,1,0,0,0,1,0},
       {0,0,1,0,1,1,1,0,0,1,1,1,0,1,1},
       {1,1,0,0,1,1,1,0,1,1,0,0,0,1,0},
       {0,0,0,0,1,1,0,1,0,0,0,0,1,0,0},
       {0,0,0,0,1,1,0,1,0,0,0,0,1,0,0},
       {0,1,1,0,1,1,1,1,1,1,1,0,0,0,0}}))
-- Glider on a 15x14 board
-- 840 distinct generation
-- loop afterward -> return -1
-- takes ~4-5 seconds on the same bad laptop
print(f({{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},
       {0,0,0,0,1,0,0,0,0,0,0,0,0,0,0},
       {0,0,0,0,0,1,0,0,0,0,0,0,0,0,0},
       {0,0,0,1,1,1,0,0,0,0,0,0,0,0,0},
       {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},
       {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},
       {0,0,0,0,0,0,0,0,0,0,0,0,0,0,0},
       {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}}))

果冻，26 25字节

ṙ-r1¤SZµ⁺_|=3
ÇÐĿ-LiṪÇ$$?

在线尝试！或验证所有测试用例。

更大的测试用例（来自@Katenkyo的答案）：15×15稳定 |15×14滑翔机

怎么运行的

ṙ-r1¤SZµ⁺_|=3  Helper link. Argument: G (grid)
               This link computes the next state of G.

    ¤          Evaluate the three links to the left as a niladic chain.
 -               Yield -1.
   1             Yield 1.
  r              Range; yield [-1, 0, 1].
ṛ              Rotate the rows of G -1, 0 and 1 units up.
     S         Compute the sum of the three resulting grids.
               Essentially, this adds the rows directly above and below each given
               row to that row.
      Z        Zip; transpose rows and columns.
       µ       Convert the preceding chain into a link and begin a new chain.
        ⁺      Apply the preceding chain once more.
               This zips back and adds the surrounding columns to each column.
         _     Subtract G from the result.
               Each cell now contains the number of lit cells that surround it.
          |    That the bitwise OR of the result and G.
               Notably, 3|0 = 3|1 = 2|1 = 3.
           =3  Compare each resulting number with 3.


ÇÐĿ-LiṪÇ$$?    Main link. Argument: G (grid)

ÇÐL            Repeatedly apply the helper link until the results are no longer
               unique. Collect all unique results in a list.
         $     Evaluate the two links to the left as a monadic chain:
        $        Evaluate the two links to the left as a monadic chain:
      Ṫ            Pop the last element of the list of grids.
       Ç           Apply the helper link to get the next iteration.
     i           Get the index of the grid to the right (the first non-unique one)
                 in the popped list of grids. This yields 0 iff the popped list
                 doesn't contain that grid, i.e., the grid reached a stable state.
          ?    If the index is non-zero:
   -             Return -1.
    L            Else, return the length of the popped list of grids.

Perl中，154个 151 144 140 137 133 129字节

包括+3 -ap0

用输入作为一行由空格分隔的数字组运行

life.pl <<< "0000 0001 0111 0010"

仅在输入立即稳定的情况下才真正需要这样做。在所有其他情况下，您也可以更方便地将其作为单独的数字行给出：

life.pl
0000
0001
0111
0010
^D

但是，对于立即稳定的配置，以这种方式提供输入将给出1而不是0。

life.pl：

#!/usr/bin/perl -ap0
map{@f=map$F[$_%@F]x3,$i-1..++$i;s%.%"0+($&+33)=~grep\$_,map{(//g)[@--1..@+]}\@f"%eeg}@Q=@F;$_=/@Q/?keys%n:$n{$_="@Q"}--||redo

几乎在这一点上击败了Mathematica ...

仅在较旧的perl版本（可以将常量用作变量）上，此126字节的解决方案才有效：

#!/usr/bin/perl -p0a
map{@f=map$F[$_++%@F]x2,-1..1;s%.%"0+($&+33)=~grep\$_,map{(//g)[@--1..@+]}\@f"%eeg}@Q=@F;$_=/@Q/?keys%n:$n{$_="@Q"}--||redo

如果确定至少有2行，则此123字节解决方案适用于所有perl版本：

#!/usr/bin/perl -p0a
@F=@F[-$#F..!s%.%"0+($&+33)=~grep\$_,map{(//g,//g)[@--1..@+]}\@F[-1..1]"%eeg]for@Q=@F;$_=/@Q/?keys%n:$n{$_="@Q"}--||redo

红宝石，207个字节

->a{r=[];(r<<a;a=(0...a.size).map{|i|(0...a[i].size).map{|j|n=0;(-1..1).map{|u|(-1..1).map{|v|n+=a[(i+u)%a.size][(j+v)%a[i].size]}};[n==3,n>2&&n<5][a[i][j]]?1:0}})while(!r.index(a));(a==r[-1])?r.index(a):-1}

我保留了每个董事会的历史记录，因此，如果我得到一个董事会，我会先知道两次发生的事情之一。首先可能是我们找到了稳定的位置，在这种情况下，它将是我们历史上最令人讨厌的事情。另一种可能性是我们有一个循环。

朱莉娅92 88字节

f(x,r...)=x∈r?(r[k=end]==x)k-1:f(sum(i->circshift(x,[i÷3,i%3]-1),0:8)-x|x.==3,r...,x)

验证

julia> f(x,r...)=x∈r?(r[k=end]==x)k-1:f(sum(i->circshift(x,[i÷3,i%3]-1),0:8)-x|x.==3,r...,x)
f (generic function with 1 method)

julia> f([0 0 0;0 1 1;0 1 0])
2

julia> f([0 0 1 1;0 1 1 1;0 1 0 0;0 1 1 1])
2

julia> f([0 1 0 0;0 1 0 0;0 1 0 0;0 0 0 0])
-1

julia> f([0 0 0 0;0 0 0 1;0 1 1 1;0 0 1 0])
4

julia> f([0 0 0 0;0 1 1 0;0 1 1 0;0 0 0 0])
0

julia> f([0 0 0 0 1 1 0 1 0 0 0 0 1 0 0;0 1 1 0 1 1 1 1 1 1 1 0 0 0 0;0 1 1 1 0 1 0 1 0 0 0 0 1 0 0;0 0 1 0 1 1 1 0 0 1 1 1 0 1 1;1 1 0 0 1 1 1 0 1 1 0 0 0 1 0;0 0 0 0 1 1 0 1 0 0 0 0 1 0 0;0 1 1 0 1 1 1 1 1 1 1 0 0 0 0;0 1 1 1 0 1 0 1 0 0 0 0 1 0 0;0 0 1 0 1 1 1 0 0 1 1 1 0 1 1;1 1 0 0 1 1 1 0 1 1 0 0 0 1 0;0 0 1 0 1 1 1 0 0 1 1 1 0 1 1;1 1 0 0 1 1 1 0 1 1 0 0 0 1 0;0 0 0 0 1 1 0 1 0 0 0 0 1 0 0;0 0 0 0 1 1 0 1 0 0 0 0 1 0 0;0 1 1 0 1 1 1 1 1 1 1 0 0 0 0])
53

julia> f([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 0 0 0 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 1 1 1 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;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 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