GAs v. GPs

I am a Ph.D. statistician (with no formal GP training) with a perfervid desire for upgrading "old" statistics,
which was formulated within the small-data setting of the day over 200 years ago, to accommodate today's big data via hybrid GP-statisitcs models.

Now, my question of long standing in my head without an answer, if you please.I was told several years ago by a University of Chicago assistant professor, who was on the board of doctoral students in GAs/GPs,
that the GA paradigm can be extended into the GP paradigm. Is this true?
If so, I would appreciate a path-of-least resistance reply from you: a reference, or a quick blurb from you.


Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Genetic Algorithms vs Genetic Programming?

If you are referring to Genetic Algorithms vs Genetic Programming, this question is probably fairly out of scope for this website; however, I'll let the admins make that judgement.

I'm not sure I understand what you mean by extending GA paradigm to GP. Someone familiar with GA can understand GP fairly quickly by recognizing that while GAs operate on bit-strings, GPs usually operate on a tree structure, usually similar to an abstract syntax tree. In other words, if a GA population consists of binary numbers: 101010001101, GP population consists of small bits of (usually) functional programs: (add 1 2), (mult (add 1 2) (sub 1 10)), ... etc.

The best introduction to GPs is bound to be one of Koza's books (all available on Amazon).

The following link is points to the most appropriate news-group and discussion: google groups


I was wondering whether this might be about genetic algorithms... I agree that these subjects are outside the scope of LtU. Only as far as genetic programming refers to representations using actual programming languages, the topic might be relevant here.