Simulated Annealing - Parameters definition

Dear all, I would be grateful if someone can give me a help in setting parameters for a Simulated Annealing optimization process. I have defined a Lower and Upper Bounds for my parameters but is there a way to specify the minimum step increase? In fact some of my parameters need to be integer (e.g. ranging from 1 to 10 with minimum step of 1) and the definition of minimum step would decrease computation time. If it's not possible do you know if there is an alternative similar optimization method where you can specify Lower, Upper Bound and minimum step increase? Thank you in advance for your suggestions and help. Regards

[Geoff Hayes]

andton - have you written the Simulated Annealing optimization code yourself or are you using some custom code from elsewhere? If the latter, please give a reference to it so that we can suggest how you might be able to integrate your requests into the code. If you have written the code yourself, you may want to attach some or all of it to your question.

Answer 1

[Alan Weiss]

If you are using simulannealbnd then there is no way to enforce integer constraints. ga is the only Global Optimization Toolbox that allows integer constraints.

Alan Weiss

MATLAB mathematical toolbox documentation

[andton]

Thanks for your hint Alan. The GA optimization now works fine but it takes too long. Do you feel to suggest any method to speed up the optimization process? I have 8 variables and 4 of them are integers and for all of them I have set up an Upper and Lower Bound. I have also reduced the tolerance to 1e-3 ('TolFun' in gaoptimset). Is there any other way to speed the optimization? i don't need a super accurate parameters estimation at this stage. Thank you in advance for your kind help. Regards