GA/EP Notes: AITGA 1.10

The last section of Chapter 1 dealt with discussing how GAs actually work. The algorithm is fairly easy to implement, but it feels like "magic" when a solution just "pops out". This section goes through some math to illustrate how GAs do what they do. I spent some time trying to understand schemas, the Schema Theorem, etc. and I kind of get it but it's still a little nebulous. Future chapters should cover these topics in detail.

I also figured out that GeneticAlgorithm<T> was working just fine. I had written a LINQ query wrong so I wasn't getting the best chromosome; I was getting the worst. Adding "descending" fixed that! Here's a plot of a simple run that shows the fitness and best solutions going up, up, and up:

* Posted at 02.06.2009 02:30:10 PM CST | Link *

Blog History