The "Principal Axis" Optimizer described here works particuarly well on my problem. Now I do know some constraints about my problem and would like to implement them to make it faster. Even though it is an unconstrained optimizer it performs better without specified constraints than the other constrained optimizers with specified constraints. Mathematica forbids to specify constraints for this particular method. Is there a way around that?

  • 2
    $\begingroup$ Adding a penalty function to the objective function is common method. I believe Mathematica often (always?) imposes constraints by such a method. $\endgroup$ – Michael E2 Apr 13 '17 at 15:22
  • $\begingroup$ Of course I can add a penalty function to introduce constraints. But I have constraints of the form $3 < x < 10$. I would like to tell the Principal Axis Optimizer that it doesn't need to check all $x \in \mathbb{R}$. With this I hope to make it more efficient at solving my problem. $\endgroup$ – tonydo Apr 27 '17 at 12:27

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.