kron_with_controls(*matrices) → numpy.ndarray¶
Computes the kronecker product of a sequence of matrices and controls.Use linalg.CONTROL_TAG to represent controls. Any entry of the outputmatrix corresponding to a situation where the control is not satisfied willbe overwritten by identity matrix elements.The control logic works by imbuing NaN with the meaning “failed to meet oneor more controls”. The normal kronecker product then spreads the per-itemNaNs to all the entries in the product that need to be replaced by identitymatrix elements. This method rewrites those NaNs. Thus CONTROL_TAG can bethe matrix [[NaN, 0], [0, 1]] or equivalently [[NaN, NaN], [NaN, 1]].Because this method re-interprets NaNs as control-failed elements, it won’tpropagate error-indicating NaNs from its input to its output in the wayyou’d otherwise expect.
Parameters: *matrices – The matrices and controls to combine with the kronecker product. Returns: The resulting matrix.