I currently have a cost function that I'm minimizing, cvxpy.norm(A@x+B@u1+G - y1)+cvxpy.norm(A@(A@x+B@u1+G) + B@u2 + G - y2), with A a Variable matrix, and everything else a Parameter object (vector or matrix, depending). CVXPY labels the second norm's curvature as being UNKNOWN (tested in isolation), but I've ensured that the argument is real (using is_real). Is there any way I can force CVXPY to treat this as convex? I always get a DCP error (objective not DCP) when solving.
Thanks!