svd issue

[Ildebrando Magnani]

Hello,

I have a 3x3 matrix A and I am trying to use svd() to compute the svd of A. I write the following code:

‌

DynamicMatrix<double, columnMajor>  U;

DynamicVector<double, columnVector> s;       // contains m singular values

DynamicMatrix<double, columnMajor>  S(3,3);     // matrix version of s, need to build it

DynamicMatrix<double, columnMajor>  V;

S = 0.;                                         // initialize singular values matrix to zero  

‌

svd( A, U, s, V);     // Computing the singular values and vectors of gqx            

‌

for ( unsigned long int i = 0; i < 3; i++){

S(i,i) = s[i];

}

cout << "--------------------------------------" << endl;

cout << A << endl;

cout << U * S * trans(V) << endl;

cout << "--------------------------------------" << endl;

‌

However, the printed outputs do not agree, but they should. How is this possible? I can’t get back the matrix A from the SVD generated by svd(). Do you have any idea of why this might be happening?

Thank you so much for your help.

‌

[Klaus Iglberger]

Hi Ildebrando!

Thanks for taking the time to report this potential defect. In this case, the problem seems to be resolved rather easily: please do not transpose the matrix V:

// ...
cout << U * S * V << endl;
// ...

Then the output for A and U * S * V are equal.

Best regards,

Klaus!

[Ildebrando Magnani]

OK, thank you very much for your help! It works!

‌

Brando