Matrix

Revision en58, by DanAlex, 2015-10-25 15:43:23

Yeah, yeah, I know you expect from me matrix jokes. What if I told you I have no jokes on that ? So, just take the blue pill and go into serious stuff like ...

Cutting to the chase

I personally find matrix multiplication as the guy who sells stolen phones at the corner of the street. I mean, you get stuff at lower price but it can break in two days and you can get busted by the cops. Or not. I really need to find better metaphors...

Matrix multiplication is a well known method. People wrote on it before quite good articles, but I think you might get stuff simpler just by looking over some problems. For the ones who are familiar with the topic, you can skip to the last two problems.

Getting high fast

Now I need to find better subtitles... However, first of all to know logarithmic matrix multiplication, you have to know logarithmic multiplication.

Basically we have to compute xn considering the multiplication operation take O(1) time. Take it straight forward we get xn = x * x * .. * x , so O(N). Let's try to reduce it step by step. Let's take xn = x2 * x2 * .... And we multiply by x if n is odd. This should work fine and the constant is reduced at half. Right... Similarly we can go to xn = xsqrt(n) * xsqrt(n) * ... and this goes to O(sqrtN). This reasoning stops here.

To get it faster you have to simply observe that xn = xn / 2 * xn / 2 for n even and xn = xn / 2 * xn / 2 * x for n odd. The two terms are the same and the third is constant, so we really need to compute xn / 2 once. And xn / 4 once. And so on. Therefore the O(logN) complexity.

Now, notice that we did not specified that x is an integer or a number. The same rules hold for other mathematical associative structures such as matrices.

Don't get stuck with struct

If you sayin' Y U NO REMEMBER MATRIX, then let me refresh your maths knowledge. You don't really need to know much about matrices to use put recurrences in a matrix multiplication form. Multiplying squared matrices is straight forward. Given two matrices their product is

Each element in each row in M is multiplied by its correspondent in the columns of N. If you find it simpler to remember, just imagine horizontal rows splitting matrix M and vertical row splitting matrix N and then match each row with each column.

Let's make a structure in which we keep a matrix. If new to matrices you should get an idea how matrix multiplication works from the code below.

struct matrix {
  // N is the size of the matrix
  int m[N][N];
  matrix()
  {
     memset(m,0,sizeof(m));
  }
  matrix operator * (matrix b)
  {
     matrix c = matrix();
     for (int i = 0; i < N; ++i)
       for (int j = 0; j < N; ++j)
         for (int k = 0; k < N; ++k) 
           c.m[i][j] = (c.m[i][j] + 1LL * m[i][k] * b.m[k][j]) % M;
     return c;
  }
  ...
};

Notice that we define the multiplication operation. We specifically did this so we can use a matrix exactly as a number in the logarithmic multiplication algorithm. So the code will be the same for an int and a matrix. Pretty cool if I do say so. And I do.

matrix modPow(matrix m,int n)
{
  if ( n == 0 )
    return unit; // the unit matrix - that is 1 for principal diagonal , otherwise 0
  matrix half = modPow(m,n/2);
  matrix out = half * half;
  if ( n % 2 )
    out = out * x;
  return out; 
}

Note that we could have defined an operator for power multiplication or used a template that we could have applied for a general type, but I find the implementation above more clear due to the clarity of the recurrence.

N-th Fibonacci term

For starters, let's define:

Fn = Fn - 1 + Fn - 2 with F1 = 1, F2 = 1

We need to put this in the form of a matrix recurrence. Well, each term is dependent of other consecutive two. This is a good clue we need just a 2 row matrix. So, from Fn - 2 and Fn - 1 we need to compute Fn. To keep the recurrences squared we will compute from the pair (Fn - 2, Fn - 1) the pair (Fn - 1, Fn).

Fn = Fn - 1 * 1 + Fn - 2 * 1 Fn - 1 = Fn - 1 * 1 + Fn - 2 * 0

Or, as matrices:

Going one step backwards we got:

Finally:

Getting at power n takes logarithmic time , so that is just... fast.

Bits and pieces

As you can see, this technique can be used to calculate the n-th term of a linear recurrence. In the following example we need to find out how many arrays of length n with maximum k consecutive 0 bits are there. ( n ≤ 109, k ≤ 40 )

Let's suppose n is small enough so we can use dynamic programming to solve the problem. Denote Dn, k = number of arrays of length n which end in k number of 0s

As you can guess one can make two moves: add a 0 and add a 1. Therefore, from state (n, k) we can go to states (n + 1, k + 1) and (n + 1, 0). So and Dn, k = Dn - 1, k.

History

 
 
 
 
Revisions
 
 
  Rev. Lang. By When Δ Comment
en83 English DanAlex 2016-03-08 03:35:36 59 Tiny change: ' & ... & 0\\ 0 && 0 & ... ' -
en82 English DanAlex 2016-03-08 03:24:58 2 Tiny change: '= D_{n-1,k} $\n\nThe' -> '= D_{n-1,k-1} $\n\nThe'
en81 English DanAlex 2016-02-07 15:39:28 6 Tiny change: ' problems.\n\n### Ge' -> ' problems. [cut]\n\n### Ge'
en80 English DanAlex 2016-01-25 16:52:59 2 Tiny change: 't = out * x;\n retur' -> 't = out * m;\n retur'
en79 English DanAlex 2015-10-27 01:46:23 8
en78 English DanAlex 2015-10-27 01:45:38 4 Tiny change: '`==` and `&&` are bric' -> '`==` and `oo` are bric'
en77 English DanAlex 2015-10-27 01:45:03 559
en76 English DanAlex 2015-10-25 16:59:30 0 (published)
en75 English DanAlex 2015-10-25 16:54:33 748
en74 English DanAlex 2015-10-25 16:45:17 427
en73 English DanAlex 2015-10-25 16:41:18 31
en72 English DanAlex 2015-10-25 16:40:30 1120
en71 English DanAlex 2015-10-25 16:22:13 1 Tiny change: '-+ \n~~~~~' -> '-+ \n~~~~'
en70 English DanAlex 2015-10-25 16:21:53 797
en69 English DanAlex 2015-10-25 16:08:48 706
en68 English DanAlex 2015-10-25 15:57:10 66
en67 English DanAlex 2015-10-25 15:55:47 97
en66 English DanAlex 2015-10-25 15:54:38 8
en65 English DanAlex 2015-10-25 15:53:27 18
en64 English DanAlex 2015-10-25 15:52:45 7
en63 English DanAlex 2015-10-25 15:52:11 2 Tiny change: '. & 0 \\ ... \\ 0 & 0' -> '. & 0 \\ .&.&. \\ 0 & 0'
en62 English DanAlex 2015-10-25 15:51:21 376
en61 English DanAlex 2015-10-25 15:46:52 32
en60 English DanAlex 2015-10-25 15:45:55 20
en59 English DanAlex 2015-10-25 15:45:10 191
en58 English DanAlex 2015-10-25 15:43:23 12 Tiny change: 'here. ( $n<=10^9 , k <= 40$ )\n\' -> 'here. ( $n \le 10^9 , k \le 40$ )\n\'
en57 English DanAlex 2015-10-25 15:42:56 36
en56 English DanAlex 2015-10-25 15:42:01 76
en55 English DanAlex 2015-10-25 15:39:58 6 Tiny change: 'm. Denote `D_{(n,k)}' = number ' -> 'm. Denote $D_{n,k}$ = number '
en54 English DanAlex 2015-10-25 15:39:35 15
en53 English DanAlex 2015-10-25 15:38:28 143
en52 English DanAlex 2015-10-25 15:35:40 5
en51 English DanAlex 2015-10-25 15:35:21 452
en50 English DanAlex 2015-10-25 03:19:21 119
en49 English DanAlex 2015-10-25 03:18:05 4 Tiny change: 'matrix} ^ (n-2) = \begin{' -> 'matrix} ^ {n-2} = \begin{'
en48 English DanAlex 2015-10-25 03:17:48 182
en47 English DanAlex 2015-10-25 03:16:40 206
en46 English DanAlex 2015-10-25 03:14:12 7 Tiny change: 'matrix} 0 1 & 1 1 \end{bm' -> 'matrix} 0 & 1 \\ 1 & 1 \end{bm'
en45 English DanAlex 2015-10-25 03:13:45 258
en44 English DanAlex 2015-10-25 03:10:44 10
en43 English DanAlex 2015-10-25 03:10:10 347
en42 English DanAlex 2015-10-25 03:06:27 156
en41 English DanAlex 2015-10-25 03:01:27 37 Tiny change: 'more clear.' -> 'more clear due to the clarity of the recurrence.'
en40 English DanAlex 2015-10-25 03:00:37 186
en39 English DanAlex 2015-10-25 02:58:28 7 Tiny change: ' c;\n }\n};\n~~~~' -> ' c;\n }\n ...\n};\n~~~~'
en38 English DanAlex 2015-10-25 02:58:09 571
en37 English DanAlex 2015-10-25 02:53:18 584
en36 English DanAlex 2015-10-25 02:45:49 44
en35 English DanAlex 2015-10-25 02:44:10 23
en34 English DanAlex 2015-10-25 02:42:28 257
en33 English DanAlex 2015-10-25 02:36:49 8 Tiny change: 'ix} b_1 & b_2 \\ b_3 & b_4 \en' -> 'ix} b_1 & | & b_2 \\ b_3 & | & b_4 \en'
en32 English DanAlex 2015-10-25 02:36:17 5 Tiny change: '& a_2 \\ ---- \\ a_3 &' -> '& a_2 \\ - & - \\ a_3 &'
en31 English DanAlex 2015-10-25 02:35:58 3 Tiny change: '& a_2 \\ - - \\ a_3 &' -> '& a_2 \\ ---- \\ a_3 &'
en30 English DanAlex 2015-10-25 02:35:41 4 Tiny change: '& a_2 \\ ----- \\ a_3 &' -> '& a_2 \\ - - \\ a_3 &'
en29 English DanAlex 2015-10-25 02:35:24 9 Tiny change: ' & a_2 \\ a_3 & ' -> ' & a_2 \\ ----- \\ a_3 & '
en28 English DanAlex 2015-10-25 02:34:35 7 Tiny change: '{bmatrix} $ = \n$\begin{bma' -> '{bmatrix} = \begin{bma'
en27 English DanAlex 2015-10-25 02:34:14 4
en26 English DanAlex 2015-10-25 02:33:38 18
en25 English DanAlex 2015-10-25 02:32:46 5 Tiny change: '{bmatrix} = \n\begin{b' -> '{bmatrix} $ = $\n\begin{b'
en24 English DanAlex 2015-10-25 02:32:15 76
en23 English DanAlex 2015-10-25 02:31:14 68
en22 English DanAlex 2015-10-25 02:30:11 217
en21 English DanAlex 2015-10-25 02:27:25 43
en20 English DanAlex 2015-10-25 02:26:39 4
en19 English DanAlex 2015-10-25 02:25:45 11
en18 English DanAlex 2015-10-25 02:24:46 7
en17 English DanAlex 2015-10-25 02:24:31 3
en16 English DanAlex 2015-10-25 02:24:09 4
en15 English DanAlex 2015-10-25 02:23:44 20
en14 English DanAlex 2015-10-25 02:22:34 75
en13 English DanAlex 2015-10-25 02:20:26 7
en12 English DanAlex 2015-10-25 02:20:08 45
en11 English DanAlex 2015-10-25 02:19:17 54
en10 English DanAlex 2015-10-25 02:17:40 73
en9 English DanAlex 2015-10-25 02:15:46 1 Tiny change: ' \n\n$M = \[a & b & c' -> ' \n\n$M = [a & b & c'
en8 English DanAlex 2015-10-25 02:15:30 56
en7 English DanAlex 2015-10-25 02:14:55 95
en6 English DanAlex 2015-10-25 02:11:32 478
en5 English DanAlex 2015-10-25 02:02:29 333
en4 English DanAlex 2015-10-25 01:57:53 6 Tiny change: ' = x^{sqrt n} * x^{sqrt n} * ...$ a' -> ' = x^{sqrt(n)} * x^{sqrt(n)} * ...$ a'
en3 English DanAlex 2015-10-25 01:57:35 10
en2 English DanAlex 2015-10-25 01:56:21 268
en1 English DanAlex 2015-10-25 01:53:13 1004 Initial revision (saved to drafts)