Genetic algorithms¶

I’m fascinated by the genetic algorithms family. Basic, sturdy genetic algorithms, John Holland style. Genetic Programming, Koza-style. Evolutionary strategies, à la Bienert, Techenberg and Schwefel. The thing thats interesting about these particular learning tools is that they require keeping a population of sub-optimal solutions about. One is less focussed on finding an optimising value than a suite of tools for solving similar problems. They can be made robust against very noisy fitness functions.

Things to learn about - Evolving “robustness”, and “modularity”, whatever those are. Implementations on massively-parallel shared-nothing architecture. Generalising to new inputs. Generalising from an evolutionary to a market metaphor for fitness, or finding some nice combination of those two metaphors.

Next I shall look at:

• E. Vladislavleva’s symbolic regression for genetic programming publications look tasty
• Bill Tozier’s Pragmatic Genetic Programming has the kind of curmudgeonly tone which always seems to hook me in to technical writing.
• Appendix A in Cosma Shalizi’s paper Dynamics of Bayesian Updating with Dependent Data and Misspecified Models. Electronic Journal of statistics. doi:10.1214/09-EJS485. (Arxiv).... maps Bayesian inference to the replicator equation (!)
• Brownlee’s “Clever algorithms” book includes some examples of GP and other nature-inspired learning techniques.
• The extensive, free, Koza-endorsed Field Guide to Genetic Programming

Thought-provoking already:

• Arthur, W. Brian. 1994. Inductive Reasoning and Bounded Rationality: The El Farol Problem. American Economic Review. (reprint.) The evolutionary dynamics of solving the minority game tickles something in my soul.
• Baum, Eric B. 2004. What Is Thought?. (see also my notes on that book)
• Holland, John H. 1992. Adaptation in Natural and Artificial Systems: An Introductory Analysis with Applications to Biology, Control, and Artificial Intelligence. (amazon.)
• Bill Tozier in general