A new strategy for adapting the mutation probability in genetic algorithms

Abstract

Traditionally in Genetic Algorithms, the mutation probability parameter maintains a constant value during the search. However, an important difficulty is to determine a priori which probability value is the best suited for a given problem. Besides, there is a growing demand for up-to-date optimization software, applicable by a non-specialist within an industrial development environment. These issues encourage us to propose an adaptive evolutionary algorithm that includes a mechanism to modify the mutation probability without external control. This process of dynamic adaptation happens while the algorithm is searching for the problem solution. This eliminates a very expensive computational phase related to the pre-tuning of the algorithmic parameters. We compare the performance of our adaptive proposal against traditional genetic algorithms with fixed parameter values in a numerical way. The empirical comparisons, over a range of NK-Landscapes instances, show that a genetic algorithm incorporating a strategy for adapting the mutation probability outperforms the same algorithm using fixed mutation rates.