New Hierarchical Structure Learning Algorithm Having the Nonstationary Random Environment to Each Level

In this paper, the hierarchical structure learning automata operating in the S-model nonstationary random environment in each level is considered, and, based on the concept of the relative reward strength algorithm, new hierarchical structure learning algorithm is constructed. It is shown that the proposed algorithm ensures convergence with probability 1 to the optimal path under a certain type of nonstationary random environment. The efficacy of the proposed algorithm is demonstrated by the computer simulation