Couple GA optimization with Nelder-Mead for curve fitting

[blakeaw]

It might be interesting to try coupling the GA-based optimization with a Hill climbing method like Nelder-Mead (simplex method) in the optimize.response_fit function. This would work by passing the optimum from the GA to scipy.optimize.minimize for further iteration to the optimum. Since both methods are pretty fast I don't think there should be a significant performance penalty, but this can be assessed.