Noise, fitness distribution, and selection intensity in genetic algorithms
Timothy Meekhof, Terence Soule · 2011
Many Genetic Algorithm (GA) problems have noisy fitness functions. In this paper, we describe a mathematical model of the noise distribution after selection and then show how this model of the noise distribution can be used to model the real, underlying selection intensity of the GA population, which promises to give us a better way to model GA convergence in the presence of noise.