please help for matrix diagonalization in c++

[henrymorgan]

ı have to prepare a project. but ı don t know how can ı do this. program will find a matrix P for a square matrix A such that P^-1 A P

[Yay295]

for A = I, P = P^-1.

You're going to need to give us more information before we can help you. Some code would be good as well.

[henrymorgan]

an algorithm for this

http://tinypic.com/view.php?pic=2lxjrt0&s=8#.U4IyS_lv7pQ

but ı have no idea how can ı write codes

[ejder]

I think the problem is the step 3. Can someone use Eigen to evaluate eigenvalues?

[Yay295]

I saw someone asking how to invert a matrix with C++ yesterday. Since two people have now asked, I am going to assume that this is a school project. We will not do your homework for you. As it is, it's not that hard to find code for this if you use Google.

[ejder]

To invert a matrix is much more easier than this. How can you come up with this idea?

[Yay295]

Ah, sorry. I just skimmed over it and thought it looked like matrix inversion. It still looks like homework though.

[henrymorgan]

I ve been working on it since last week and I couldn't find a way to do it. this program requires some math, ı hope someone can help me because ı m really tired of this

[Yay295]

Alright, first we need to get the eigenvalues.
Using those we can get the eigenvectors.
We can put those together to form P and P^-1.
and finally we can solve for D.

Do you want this to work with any values, or can we stick to ints?

[henrymorgan]

I used double, at first. However lets use ' int '.

[henrymorgan]

hey Yay295, are you working on it or have you gone?

[Yay295]

I did go eat lunch, but I'm here. Have you gotten anything done yourself?

[henrymorgan]

Nope, sorry :)

[Yay295]

This is too much thinking for me right now. I'll leave you with what I've done though. It doesn't get the correct eigenvalues.

#include <iostream>
#include <vector>
#include <cmath>

typedef int unit;

void printMatrix ( const std::vector<unit> & matrix, int width, int height );

void eigenvalues ( std::vector<unit> matrix, std::vector<unit> & eValues, std::vector<unit> & found, unit multiplier = 1 );

int main ( )
{
	const int size = 2;

	std::vector<unit> matrix ( size * size, 0 );
	std::vector<unit> eValues ( size, 0 );
	std::vector<unit> found ( eValues );

	eigenvalues ( matrix, eValues, found );

	printMatrix ( matrix, size, size );

	printMatrix ( eValues, size, 1 );

	printMatrix ( found, size, 1 );
}

void printMatrix ( const std::vector<unit> & matrix, int width, int height )
{
	for ( int i = 0; i < height; ++i )
	{
		for ( int j = 0; j < width; ++j )
			std::cout << matrix[j+i*height] << ' ';

		std::cout << '\n';
	}

	std::cout << '\n';
}

void eigenvalues ( std::vector<unit> matrix, std::vector<unit> & eValues, std::vector<unit> & found, unit multiplier )
{
	int mSize = matrix.size ( );

	if ( mSize > 4 )
	{
		int width = sqrt ( mSize );

		for ( int i = 0; i < width; ++i ) // for every item in the top row
		{
			std::vector<unit> temp;

			int newSize = ( width - 1 ) * ( width - 1 );

			for ( int j = width; j < mSize; ++j ) // start at the second row
			{
				if ( !( j % ( width + i ) ) ) ++j;

				else temp.push_back ( matrix[j] );
			}

			eigenvalues ( temp, eValues, found, matrix[i] );
		}
	}

	else
	{
		unit A = matrix[0], B = matrix[1], C = matrix[2], D = matrix[3];
		
		unit vOne = -1*B + sqrt ( B*B - 4*A*C ) / 2*A, vTwo = -1*B - sqrt ( B*B - 4*A*C ) / 2*A;

		vOne *= multiplier; vTwo *= multiplier;

		int fSize = found.size ( );

		for ( int i = 0; i < fSize; ++i )
		{
			if ( !found[i] )
			{
				found[i] = 1;
				eValues[i] = vOne;
				found[i+1] = 1;
				eValues[i+1] = vTwo;
				break;
			}
		}
	}
}

[henrymorgan]

First of all, thank you for the effort that you've spent but how can we enter the info of dimension ?

[henrymorgan]

I'm stuck :) Does anyone have any ideas?