I have some linear equations:

( (v1.*x) + (v2.*y) ) * A = (v3.*z) * B

( (w1.*x) + (w2.*y) ) * A = (w3.*z) * B

where

v1, v2, v3, w1, w2 and w3 are known row vectors (with complex numbers)

A and B are matrices, usually with more columns than rows. (real)

x, y, z are row vectors; x known, y and z unknown.

I seek to solve for y and z.

What is the appropriate way to use Octave to solve these?

In truth each side represents a continuous function on [0,2pi] with

respect to different sets of orthogonal basis functions. The matrices A

and B are the Fourier Coefficients of these orthogonal functions, and I

can calculate as many or few as I feel inclined.

Thanks for any advice,

John

[hidden email]