Optimization with Probabilistic Constraint

Hi All,

I plan to solve an optimization problem using the genetic algorithm or fmincon. Nonlinear constraints are the probabilities of failure calculated by the Monte Carlo method. But I don't know how to define this constraints so that the optimization can be done properly.

Do you know a sample code, link, or document that model such constraints and explains how to implement it?

Thank you for guiding me.

Answer 1

[John D'Errico]

You cannot use a tool like fmincon for this. It requires deterministic and differentiable constraints. For that matter, even GA probably prefers repeatability in the constraint and objective functions, that the constraint does not randomly change.

I suppose GA will likely survive it, returning something reasonable despite the fuzzy boundary. To try it with fmincon would be a complete waste of time though.

As for a link that shows how to do it, you are the only one I've ever seen who wants to do this.