How to get the best results from gamultiobj for large number of objectives ... say 12 objectives to be maximized ?

Hi, I am using gamultiobj to maximize the output power of 12 wind turbines, (individually), and comparing that with maximizing the total power instead. I become able to modify the options for GA so the results look logic and reasonable. BUT for gamultiobj: the run always terminated because the number of generations are exceeded, and the results don't look good enough, I think the solver should be more stronger in search, the number of generations should be larger, or the population size should be increased, but I don't how to do that from the documentations ... http://www.mathworks.com/help/gads/examples/multiobjective-genetic-algorithm-options.html

How can I get the best possible results when maximizing say 12 objectives individually like my case ? by the way, I don't have any constrains at all thanks in advance

Answer 1

[Alan Weiss]

You can adjust the number of generations using the Generations option in gaoptimset. You can adjust the population size using the PopulationSize option. Be sure to pass your options to the solver; I mean, make the call

 x = gamultiobj(FITNESSFCN,NVARS,A,b,Aeq,beq,LB,UB,nonlcon,options)

Since you don't have constraints, put [] for A, b, Aeq, beq, LB, UB, and nonlcon. Your call will look like this:

 x = gamultiobj(FITNESSFCN,NVARS,[],[],[],[],[],[],[],options)

Alan Weiss

MATLAB mathematical toolbox documentation

[Mamdouh]

1- thanks, regards, and respect dear Alan for concern and reply, your answers always and really clear and help, 2- did it mention anywhere that the ga options can be used in gamultiobj ? if not, please try to add it explicitly in the documentations. 3- after adjusting the options (increasing the population size and the number of generations, etc.) do you see gamultiobj can handle 12 objective functiones efficiently ?

[Alan Weiss]

I'm glad that you find my answers helpful. I see no issue in using gamultiobj on a 12-dimensional problem, except that it is often difficult to analyze or plot the answer.

Alan Weiss

MATLAB mathematical toolbox documentation