next up previous print clean
Next: Other operators Up: Basic Operators Previous: Crossover

Mutation

The operator responsible for introducing new genetic material into the population is called mutation (or more precisely jump mutation). With binary encoding, this operator works by randomly selecting a bit and then flipping it. In general, mutation is applied with a relatively low probability, usually less than 0.1. The idea is that although mutations are critical to prevent the population from loosing their genetic diversity (that is, mutations force the algorithm to look into other areas of the search space) they are potentially disruptive causing the algorithm to lose information from a promising search area.


next up previous print clean
Next: Other operators Up: Basic Operators Previous: Crossover
Stanford Exploration Project
11/11/2002