07-02-2013 02:08 PM
Does anyone know what algorithm is used for the "Positive Definite" option of the Inverse Matrix VI? I'm trying to reimplement some LabVIEW code in C (no, this comment is not intended to start a flame war, or to incite responses of "why don't you just use G with such and such?" I have my reasons. They are good. Drop it.) and I'm getting results that are similar, but not identical. All I've managed to find so far in the wide world of web is one sentence suggesting that Cholesky decomposition is used, but is that so? Or is it LDL^T? Or something else?
Thanks,
Jason
07-03-2013 05:12 PM
Hello Jason_S,
If you double click subVI, you can view its block diagram. If you do this for Inverse Matrix.vi, then select the "Positive Definite" case in the case structure, you will see that it calls a subVI called Real Inverse Matrix (Cholesky).vi. This would suggest that the Cholesky Decomposition algorithm is the one that is used. However, since Real Inverse Matrix (Cholesky).vi calls a library function, you cannot determine the exact algorithm that is used.