GA/EP Notes: AITGA 3.2 to 3.3
Section 3.2 discusses modeling sexual selection using GAs. Specifically, Mitchell discusses work that was done to see if a specific trait prefered by females would be expressed in males even if that trait was harmful to survival to the male. The work showed that the trait would always spread among males even with its harmful tendency. 3.3 talked about modeling ecosystems, which targets modeling systems with interactions between organisms. Mitchell talked about a tool called Echo, which allowed the modelers to look at deception, mating, etc. via agents. She emphasized that there is a tradeoff between modeling a system that is too complex (which would be no different that observing a real-world system) and modeling a system that is too simple (which wouldn't aid the modeler with intuitive results).
I'm still working on ExpressionEvolver. I've been able to work out a number of bugs and I've made some changes after studying Koza's work in a little more detail. Today I'm focusing on adding mutation, more operators, and multi-parameter methods. This is fun stuff!
* Posted at 02.17.2009 08:20:08 AM CST | Link *