Can cvx solve quadratic programming problems? (Moderator note: this particular quadratic programming problem is non-convex)

Nonconvex
1

For example:

cvx_begin
variables x y
minimize x*y
subject to
x>=0.5;
y>=0.5;
x<=1;
y<=1;
cvx_end

and it says “Invalid quadratic form(s): not a square.” Is cvx only applicable for squared objective functions like x^2?

2

No, it is not applicable for arbitrary, nonconvex quadratic problems, like yours.

It is applicable for quadratic problems which have a convexity proof which fits the DCP rules, for example by using x^2, quad_form, norm, sum_squares, quad_over_lin and maybe more.

1 Like
3

Further elaboration for @Michal_Adamaszek 's reply is at