gamultiobj - If I want that x(34)=x(23)*x(46) which function I must customize?

I want that x(34)=x(23)*x(46) in each element of population considered for evaluating the objective functions. Which file I must customize?

(it's a strange request, but it's a long story).

Answer 1

[Image Analyst]

I'd guess that you'd customize the file that contains the x data. Or else customize the m-file so that it changes one of those three elements of x to force the equality to hold. You're right - it is a strange request.

[Gabboc]

Sorry, I forgot to say you that the question is relative to Multiobjective Optimization, then relative to gamultiobj. My question is if I must customize creationfcn or....???

[Image Analyst]

That doesn't help. Read this first, before you post again: http://www.mathworks.com/matlabcentral/answers/6200-tutorial-how-to-ask-a-question-on-answers-and-get-a-fast-answer

[Gabboc]

I explain you my entire problem hoping you can help me. I have 10 functions to maximize(30 variables). These functions are regression (stepwise) functions like this y(1)=b*x(1)+c*x(2)+...+d*x(1)*x(2)+...

I want to minimize my functions, limiting it: y(1)<t , y(2)<o .....

Sadly, I don't know how constrain the functions due to x(1)*x(2) terms because in gamultiobj constraint are expressed in matricial form.

I'm thinking to add variables to my problem trasforming my fitness function in x(4)=x(1)*x(3); y(1)=b*x(1)+c*x(2)+...+d*x(4)+.. ;

and the constraint is A=[b c ...d]; b=t;

but the problem is that the population that verify the constraint don't consider (logically) that x(4)=x(1)*x(3) so I'm looking for a way to gengerating and using in gamultiobj only population that x(4)=x(1)*x(3);

[Image Analyst]

Sorry, I'm not an optimization expert.