Need help with writing the objective function for a genetic algorithm

1 次查看（过去 30 天）

Ameera 2014-7-23

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/143082-need-help-with-writing-the-objective-function-for-a-genetic-algorithm

编辑： Ameera 2014-7-24

I usually use AMPL to write mixed integer programs, but now I need to implement a Genetic Algorithm so I'm trying to write my objective function on Matlab. I cannot handle the summation over specific ranges like I used to do in AMPL. It'll be great if someone can help me figuring out how to write the attached objective function in Matlab. Thanks.

2 个评论
显示无隐藏无

José-Luis 2014-7-23

No attached file. Please show what you have tried so far.

Ameera 2014-7-24

编辑：Ameera 2014-7-24

在 MATLAB Online 中打开

I have written the fitness function copied below, it is a simplified function of what I'm trying to do. My issue now is with the GA optimization tool:

1-My only variable is "x" (2x3 matrix in the given example below), what should be the number of variables in this case? Is it 1 or 6?

2-How to specify that x is binary?

3-I have the following constraint in my problem: sum(t in 1..T) x(i,t) <=1, for every i. How can I enter that constraint in the linear inequalities boxes.

Here is my fitness function:

function [x,fval] = APP(alpha,k,r,v,vz,x0)
[x,fval] = fminunc(@APPnest,x0);
function obj = APPnest(x)
obj = 0;
for t = 1:3
    for i = 1:2
        z = 0;
        for u = 1:t
            z = z + k(i,t-u+1) * x(i,u);
        end
        nem = v(i)*z;
        w = 0;
        for j = 1:2
            for u = 1:t
                w = w + v(j) * k(j,t-u+1) * x(j,u);
            end
            den = vz + w;
        end
        obj = obj + alpha(t)*r(i)*(nem/den);
    end
end
obj = -obj;
end
end