Hello, I'm currently working on an optimization problem using your library and the COBYLA algorithm. My objective function is concave and a maximum is well defined.
I see you have a method, nlopt_set_initial_step or nlopt_set_initial_step1, which you can use to set the initial step size. I'm wondering though, is there a way to set a fixed step size? My problem's variables are only (logically) able to be WHOLE numbers, so all of the optimization iterations that change my variables by very small fractions of a whole are a waste of time. Moreover, my problem contains a large amount of variables and constraints, so it would be ideal to eliminate these wasted iterations. I noticed that there is a method, nlopt_set_default_initial_step, that is not mentioned in the online reference. I'm not sure if this is the answer I'm looking for, the name implies that it too only sets the initial step size. I greatly appreciate any help and hope I haven't just overlooked something in the tutorial or reference. Regards, Adam O. _______________________________________________ NLopt-discuss mailing list [email protected] http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/nlopt-discuss
