Genetic algorithm : Best score is not equal to fitness

Hello,

I am currently using a genetical algorithm, and using a output fuction, I display the best state.Score value at each iterations.

Once I stop the algorithm, I get my element (its genome) with my best score. The thing is, if I use my fitnessfunction on that element, outside the genetic algorithm framework (so, using something like FitnessFunction(element)) . I get a value that is slightly different from the state.Score, even though this element is the best one.

I would like to know if anyone knows how can that difference be, is it matlab defining the score using another definition than the fitness ?

Thank you

[Walter Roberson]

Check to see if there is a small constraint violation at that position. Constraint violations introduce a penalty on the score.

[Geoff Hayes]

Boris - also, what does slightly different mean?