SISO Behavior when multiplied with a matrix

Question

SISO Behavior when multiplied with a matrix

ilayn opened this issue 9 years ago · comments

When a SISO Transfer() or State() multiplied with a p x m matrix, should it reject due to size mismatch or elementwise multiply with each element of the matrix P as if an overloaded Kronecker Product kron(P,G) ?

Ilhan Polat · Answer 1 · Thu Jun 30 2016 19:57:33 GMT+0800 (China Standard Time)

Currently rejected if systems are multiplied from the right. But for the left multiplication numpy tries element by element method and each entry of an array is allowed for SISO systems. The result becomes a numpy array with object dtype. Currently I don't know how to prevent this.

Ilhan Polat · Answer 2 · Tue Sep 20 2016 17:56:15 GMT+0800 (China Standard Time)

Since the api for matrix multiplication requires @ or dot(), and star is elementwise mult, right multiplication is also allowed now.

Ilhan Polat · Answer 3 · Wed Feb 01 2017 16:30:31 GMT+0800 (China Standard Time)

I've asked about this on SO but so far didn't receive any solution.

Ilhan Polat · Answer 4 · Tue Jun 20 2017 01:09:39 GMT+0800 (China Standard Time)

Thanks to the fine people of NumPy mailing list, I'll set __array_ufunc__ attribute on the representation classes to override ndarray.__mul__().

Finally...

Ilhan Polat · Answer 5 · Tue Jun 20 2017 01:11:59 GMT+0800 (China Standard Time)

this will require NumPy 1.13+ though I don't expect any objections anyways