Chapter 3. Selection Strategies & Elitism

Table of Contents

Truncation Selection
Fitness-Proportionate Selection
Roulette Wheel Selection
Stochastic Universal Sampling
Rank Selection
Tournament Selection
Sigma Scaling

Selection is an important part of an evolutionary algorithm. Without selection directing the algorithm towards fitter solutions there would be no progress. Selection must favour fitter candidates over weaker candidates but beyond that there are no fixed rules. Furthermore, there is no one strategy that is best for all problems. Some strategies result in fast convergence, others will tend to produce a more thorough exploration of the search space. An evolutionary algorithm that appears ineffective with one selection strategy may be transformed by switching to a strategy with different characteristics. This chapter describes the most commonly used selection strategies (all of these strategies are supported in the Watchmaker Framework for Evolutionary Computation via different implementations of the SelectionStrategy interface).

Truncation Selection

Truncation selection is the simplest and arguably least useful selection strategy. Truncation selection simply retains the fittest x% of the population. These fittest individuals are duplicated so that the population size is maintained. For example, we might select the fittest 25% from a population of 100 individuals. In this case we would create four copies of each of the 25 candidates in order to maintain a population of 100 individuals. This is an easy selection strategy to implement but it can result in premature convergence as less fit candidates are ruthlessly culled without being given the opportunity to evolve into something better. Nevertheless, truncation selection can be an effective strategy for certain problems.