Nick Nichols

Clojure Developer and Dungeon Master

Selection

If Fitness Functions are the probability model that an organism can survive long enough to reproduce, then Selection is the use of that model. By using the fitness scores we assigned in the prior phase, we have a great tool to sample our solution pool. For example, a solution with a very high value and little weight is more important to hold on to a bag full of hammers. In most implementations, each individual is selected fairly with a probability of their fitness score against the aggregate fitness score of the entire population.

It’s as easy as that.

Genetic Algorithm Phases:

  1. Initialization
  2. Fitness Evaluation
  3. Selection
  4. Reproduction
  5. Mutation
  6. Generation Advancement and Termination