I looked at your two matrices of eigenvectors. Thank you for sending them.
Upon my investigation, I found that the column vectors (the eigenvectors) which look different are actually scalar multiples of each other. However, since you are dealing with complex eigenvectors, the multiple is not +/-1, but is rather a phase, exp(i*phi). I suggest you form ratios of each of the column pairs (column i from C++ and column i from Python) and observe that the ratio is the same for all elements. That says each column vector is just a scalar multiple of the other one, and both are valid eigenvectors. In fact, they are the same set of eigenvectors, except they can differ by a phase factor. Best regards, Stuart Brorson Northeastern University Boston, Mass ________________________________ From: Jiasen Guo <jiasenguo1...@gmail.com> Sent: Friday, January 5, 2024 10:59 AM To: Brorson, Stuart <s.bror...@northeastern.edu> Subject: Re: Help-gsl Digest, Vol 229, Issue 1 Hi, Thanks for the reply. The eigenvectors are not just differed by a sign actually. I am attaching the two eigenvector matrixes here. Thanks for your kind help. BTW, does my reply to a specific person open to all the subscribers? Best, Jiasen r
