Multi-objective optimization using discrete set of points defining the objective functions

[a...@simerics.com]

Hello all,

I noticed that the examples included with the inspyred module all involve defining objective functions in the form of polynomials. How would one go about using discrete points across the design space (obtained through some sort of DOE) in place of objective functions? Do I need to construct a response surface (or make use of some form of function fitting) in order to generate these objective functions, or is there a way to use these discrete directly in the optimization algorithms? (I'm mainly looking at PSO and NSGA-II). 

Any help regarding this would be highly appreciated.

Thanks,

Anup Doddamane

[Aaron Garrett]

I apologize. I don't understand the scenario you're describing, but you don't have to have a polynomial. I very regularly have an objective function that is the output of some simulation. If you have some specifics that you can share, I can try to help you. If you need to share it privately, use aaron.le...@gmail.com.

--
You received this message because you are subscribed to the Google Groups "Inspyred" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inspyred+u...@googlegroups.com.
To post to this group, send email to insp...@googlegroups.com.
Visit this group at https://groups.google.com/group/inspyred.
For more options, visit https://groups.google.com/d/optout.

--

Aaron Garrett, Ph.D.

Assistant Professor

Computer Science

Wofford College

429 North Church Street

Spartanburg, SC 29303

garr...@wofford.edu

(864) 597-4529