The Research of Q Learning-Based Estimation of Distribution Algorithm

This paper focuses on the theory of estimation of distribution algorithms. First, elaborated the idea of estimation of distribution algorithms, And then for the limitations of solving complex optimization problems, proposed Q Learning-Based Estimation of Distribution Algorithm. The Q learning algorithm is introduced into evolutionary computation, through the Agent and group interaction, to achieve a probability model of adaptive updates. Test functions using six classical comparative experiment, the results show that the algorithm performance is stable, running time is short, with a strong global search ability, is an efficient solving algorithm for function optimization problems.