What characteristic of a problem decides whether or not it should be solved with GP vs. GA techniques?

[Tony Stark]

What characteristic of a problem decides whether or not it should be solved with GP vs. GA techniques?

[Christian Gagné]

With GP you are usually evolving models described by a complex variable length data structure (e.g. trees) optimizing a function. With GA you are evolving values to optimize a given function. Of course you can define a parametric model using GA to produce a model, but that's often less natural than with using GP for that.

Christian

-- 

Christian Gagné

http://vision.gel.ulaval.ca/~cgagne

Le Jan 14, 2015 à 19:12, Tony Stark <34fe...@gmail.com> a écrit :

What characteristic of a problem decides whether or not it should be solved with GP vs. GA techniques?

--
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.

[Tony Stark]

1. So with GA you already have a function and you are just evolving constants?
2. With GP you just have sample inputs and outputs and you use GP to find the function, correct?

[Christian Gagné]

Your examples that are capturing many applications of these approaches. In can synthesis your two cases as cases of optimization (case 1) and learning (case 2).

However, it is not limited to that, you can, for example, use GP to evolve shape of antennas (then, it is more an optimization problem), or use GA to determine the decisions an agent should given environment (a learning problem).

So, in brief, you seem to get the point, but don't consider this as absolute.

Christian