A Comparative study of the Genetic Operators Single Point Crossover (SPX) and Uniform Crossover UX applied to the Optimization problem PathFinder

Abstract:

In this paper we have carried out an experimental study of two variants of an evolutionary algorithm applied to exploration of a maze (PathFinder) to allows to a robot find the best way out from its location within. The first variant applies the Genetic Operator Single Point Crossover (SPX) and the second the Uniform Crossver. The test mazes are formed by NxM positions with 20% of obstacles for the robot, the representation of the solution (set of steps) is based on internal absolute coordinates (North, South, East and West), but we have added some relatives coordinates (Move, Rotate Right and Rotate - Left) with the aim of perform an additional analysis. To confirm the results we have identified if exist or not statistically significant differences between the results provided by these two operators. The obtained results indicate that the best variant of the evolutionary algorithm to implement is which uses the UX crossover …

Año de publicación:

2016

Keywords:

Fuente:

google