something about understanding cma-es (beginner)

[David Schuffenhauer]

Hello everyone!

I Need some help with understanding how cma-es works: as far as I understand the covariance matrix shows the pairwise linear dependencies between the parameters. But what happens if two parameters are related in another way (lets say quadratic)? what does the cm do then?

Regards

David

[François-Michel De Rainville]

It will estimate the local relation using linear dependency. Since the covariance is estimated every step and the searched region also shrinks every step, this is often well enough.

If you really want to see what happens, you can use this tutorial and plot the covariance on a quadratic function and see the parameters adapt while searching.

Best,

François-Michel

--
You received this message because you are subscribed to the Google Groups "deap-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to deap-users+...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[David Schuffenhauer]

Thanks a lot!

[François-Michel De Rainville]

Looks like my link didn't copied, for reference I was talking about this tutorial

http://deap.readthedocs.io/en/master/examples/cmaes_plotting.html

2016-06-27 3:58 GMT-04:00 David Schuffenhauer <schuf...@gmail.com>:

Thanks a lot!

[François-Michel De Rainville]

I made a mistake explaining CMA-ES, the covariance does not model the objective function, but maximizes the probability of producing successful solution.