Constraint definition among the variables for Genetic Algorithm (GA) Optimization

Hello Everyone,

I'm trying to use multiobjective genetic algorithm to identify various parameters for my models. I have a problem in defining the constraints between the variables. If v1,v2,v3 etc.,are the variables, then I want to define constraints like v1>v2>v3...

Could anyone help me out with this?

Thanks and Regards, Karthik

Answer 1

[Alan Weiss]

The syntax for gamultiobj includes linear inequality constraints: X = gamultiobj(FITNESSFCN,NVARS,A,b) If v is your vector of n decision variables, and you want v(1) > v(2), v(2) > v(3),...,v(n-1) > v(n), set your A matrix as

 v = zeros(1,n);
 v(1) = -1;v(2) = 1;
 A = gallery('circul',v);
 A(n,:) = [];
 b = zeros(n-1,1);

Answer 2

[Karthik Vemireddy]

Thank you Alan.....

Meanwhile I have found another simpler solution. It's described in the following link in detail...

http://www.mathworks.com/help/toolbox/optim/ug/brhkghv-11.html#brhkghv-14