python 矩阵元素平方_使用python为非平方矩阵A求解Ax = b

I'm focusing on the special case where A is a n x d matrix (where k < d) representing an orthogonal basis for a subspace of R^d and b is known to be inside the subspace.

I thought of using the tools provided with numpy, however they only work with square matrices. I had the approach of filling in the matrix with some linearly independent vectors to "square" it and then solve, but I could not figure out how to choose those vectors so that they will be linearly independent of the basis vectors, plus I think it's not the only approach and I'm missing something that can make this easier.

is there indeed a simpler approach than the one I mentioned? if not, how indeed do I choose those vectors that would complete Ainto a square matrix?

解决方案

As you mentioned, np.linalg.solve needs a full-rank square matrix.

For all the other linear cases, if you are interested in min||Ax-b||^2. (you probably are) you can use np.linalg.lstsq.

Solves the equation a x = b by computing a vector x that minimizes the Euclidean 2-norm || b - a x ||^2.

The equation may be under-, well-, or over- determined (i.e., the number of linearly independent rows of a can be less than, equal to, or greater than its number of linearly independent columns). If a is square and of full rank, then x (but for round-off error) is the “exact” solution of the equation.

(bold annotation by me)

This is also mentioned in the docs of the original np.linalg.solve:

a must be square and of full-rank, i.e., all rows (or, equivalently, columns) must be linearly independent; if either is not true, use lstsq for the least-squares best “solution” of the system/equation.