Parameter uncertainties after a GA optimization

Hi everyone,

I do a genetic algorithm optimization in a mixed integer problem (e.g. 2 parameters are integers and 5 are reals). Is there a way to get an estimation of the parameter uncertainties from the ga optimization function?

Thanks,

Razvan

[Matt J]

Sounds unlikely, if you have a very complicated objective function. At the very least, you would need to know the uncertainty in your input data.

[Razvan]

What if the objective function is not very complicated, and I know the uncertainties of the data? How can I get then the parameter uncertainties after a ga optimization?

Answer 1

[Matt J]

There's a section of this FEX submission which talks about deriving uncertainty info on estimated parameters

http://www.mathworks.com/matlabcentral/fileexchange/8553-optimization-tips-and-tricks

[Jonas Bruns]

Lie!

[Matt J]

Its not a lie. Sections 29 and 30.