The picture is one of the constraints in my convex optimization (cvx) problem. ‘t’ and ‘Q’ are the optimization variable. However, when I run the code, the following error message appears: ‘Cannot perform the operation: {convex} ./ {convex}.’ I know the issue lies on the left side of the constraint, but I haven’t come up with a proper way to reformulate this inequality. Please provide your suggestions if possible.
Please show us your proof the optimization problem is convex. You haven’t shown the whole problem, so readers can’t know. However, what you do show is a linear fractional form. Is the denominator always positive?
Perhaps this can be handled by techniques along the lines of section 4.3.2 “Linear-fractional programming” of Convex Optimization – Boyd and Vandenberghe, or the problem is quasi-convex and can be handled by the bisection algorithm in section 4.2.5 “Quasiconvex optimization” of that book.
However, perhaps you are not telling us everything? If Q and t were the only optimization variables, and everything else was input data, I believe you would have gotten the error message, {affine}/{affine}, not {convex}/{convex}. So that may put my teh applicability of my preceding paragraph into greater doubt.
I check the cvx problem and code again. The error message should be{affine}/{affine} as you said. And the denominator is always positive. Q and t are the only optimization variables. I will continue to try to solve the problem. Truly grateful for your help. And please tell me if you have any other suggestions.