A learning strategy for multi-robot based on probabilistic evolutionary algorithm

Estimation of distribution algorithm (EDA) is a new evolutionary computation method based on probabilistic theory. EDA can select optimal individuals through estimating probability distribution function of a population. The capture problem among multi software robots can be solved by EDA. The capture problem involves that some pursuers pursue several evaders through part of trajectory. The trajectory was produced by the evaders during their two-dimensional random mobility. The pursuers estimate the evaders´ mobility functions and adjust their pursuit models to capture the evaders as fast as possible. The probabilistic evolutionary courses of multi-robot experiencing some competitions are analyzed in performances. The analysis shows that capture problem of multi-robot solved by EDA is better than other methods in several aspects.