I'm writing an evolution algorithm program to find best fit curves for random data sets.

This picture should give you an idea of what i'm doing...


It shows a randomised set of test data that is based on a unknown 3rd order polynomial. The red line is what I want my algorithm to generate, using a genetic algorithm.

So i have something to aim for, how long typically would it take a decent traditional algorithm to find a best fit curve for such data (given 1000 points of test data). I plan to support up to 11th order polynomials! Are we talking minutes? Seconds? Milliseconds? Hours?

I know this is a bit vague, but I'm hoping someone has some experience in this area and could impart some wisdom.