I can think of several different things that you might intend.
- If your function expects an array of trial model parameters, then you must instead expect to be passed a vector of trial model parameters, but your function can immediately reshape that into an array for internal use. The output from ga() would be a vector of model parameters, so you might need to reshape() it again afterwards for whatever presentation you want to do.
- If your function expects a scalar or vector of trial model parameters, but also needs an array of other data (for example some kind of look-up table), then see http://www.mathworks.com/help/matlab/math/parameterizing-functions.html
- Perhaps you have a function of one variable and you would like to pass in a number of trial values and have the genetic algorithm figure out which is optimal. If so then this is not a suitable task for genetic algorithms: the function is only passed in one set of model parameters at a time (unless 'UseVectorized' is true in the options)
