In this issue how can I define objective function? On the other hand does anybody know any criteria for dividing clusters that would be useful for genetic algorithm? Actually this is a framework in attachment file.
If you have not looked into it yet, a good place to start is using kmeans and Fuzzy cmeans clustering to get you a better insight into a desirable objective function.. Through MATLAB, these are simple commands which minimize the least square of the euclidean distances between cluster center and points belonging to that specific cluster.