Evolution Demos

Some toys an AI and I put together for my Practical Evolution course.

• Evolution is algorithmic

  Suppose a population lives in a world of tiles separated by trenches. Mothers can maximize their fecundity by producing many small offspring. But larger offspring are better dispersers, and have a better chance of getting from one tile to another. We express the life cycle of this population with pseudocode. We then iterate that life cycle and watch the population evolve a mean offspring size that jointly optimizes fecundity and dispersal ability. No one steering the ship. Just natural selection doing its thing.

• Same evolution, three algorithms

  Compare and contrast a selective sweep under three frameworks -- an individual-based model, an allele-frequency model (classical population genetics), and a phenotypic variance model (classical quantitative genetics).

• Breaking Hardy Weinberg

  The Hardy-Weinberg equation gives us a null model for population genetics. It tells us that if certain assumptions are met, nothing ever changes. See what happens when we break those assumptions!