how to use variables in fitness function in the genetic algorithm optimization?

i try to run genetic algorithm optimization using a fitness function that includes variables and i get this error.

--The following error occurred converting from sym to double:

DOUBLE cannot convert the input expression into a double array.

If the input expression contains a symbolic variable, use VPA.--

i used syms when i calculate the fitness function.

anyone has any ideas about this problem? thank you.

Answer 1

[Alan Weiss]

Global Optimization Toolbox function expect variables of type double, not symbolic. To use symbolic variables, first convert them to double type. For an efficient way to do this, see this example.

Alan Weiss

MATLAB mathematical toolbox documentation

[Ekin Ceyda Cetin]

Thank you, it was very helpful!

[Ekin Ceyda Cetin]

The matlabFunction I use as an objective function in fmincon does not work for ga.

'Index exceeds matrix dimensions.'

I didn't define any derivatives inside the objective function, is this related with the error? How can I solve this problem?

[Walter Roberson]

Does the function involve multiple variables? If it does then when you call matlabFunction then provide the 'vars' option as a column vector of variable names instead of using the more common row vector of variable names.

[Ekin Ceyda Cetin]

I saw this now. http://www.mathworks.com/matlabcentral/answers/113163-how-can-i-convert-symbolic-equation-to-fitness-function-which-can-be-evaluated-by-genetic-algorithm?s_tid=answers_rc1-1_p1_Topic

Thank you~~