1
$\begingroup$

I am trying to solve a non convex optimization problem with the help of sequential quadric programing. My optimization variables are complex and I have expressions for gradients, hessian etc but all these are in complex form. Let's say I am solving this problem using quadprog toolbox and inputs to this toolbox takes only real values. So what should I do about this

$\endgroup$

1 Answer 1

1
$\begingroup$

If x is the optimization complex vector variable, instead make the real variables, x_real and x_imaginary, the variables. Form the original complex variable problem in terms of those. This formulation only involves real variables.

If you directly use quadproog, rather than a more flexible front end such as YALMIP, you will have to "stack" x_real and x_imaginary into one vector variable. if you use YALMIP, you can specify the problem directly in terms of the complex variables, and YALMIP will do this transformation for you under the hood, saving you rime and reducing error propensity.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.