Using MATLAB GA package for a experimental problem

I want to use genetic algorithm toolbox in MATLAB to solve a problem. However, the objective function is not a mathematical equation that I code it and I do not know how to do it. I would like to:

1. initialize GA with a specific population 2. pass the scores of each individual to GA in a matrix format (HERE is the question, how?) 3. GA return me the next generation 4. I do the experiment on the new generation and evaluate them 5. pass the new scores to step 2 continue to reach convergence Thanks for your help!

[Walter Roberson]

What happens when you attempt to call your objective function?

[David Barry]

Please upload your code and error message

[Seyedmohamad Moosavi]

Ok, I think I have not been clear enough. There is no error. I want to use genetic algorithm toolbox in MATLAB to solve a problem. However, the objective function is not a matematical equation that I code it. I would like to: 1. initialize GA with a specific population 2. pass the scores of each individual to GA in a matrix format (HERE is the question, how?) 3. GA return me the next generation 4. I do the experiment on the new generation and evaluate them 5. pass the new scores to step 3 6. continue till I see convergence.

Sorry for not being clear.

Answer 1

[Walter Roberson]

When you ask for integral constraints, ga allows uses its own crossover and mutation and initial population functions; it does that so that it can enforce that only integers can be created in the appropriate slots. Providing initial scores is not supported by those built-in integer constraint functions.

In order to combine integer constraints with initial scores, what you need to do is to do is to not tell ga that those positions are to be treated as integers, and to instead provide your own population and crossover and mutation functions that happen to never generate non-integers in the appropriate positions.