抄録
The purpose of this paper is to construct a new learning algorithm for variable hierarchical structure learning automata operating in a nonstationary random environment by extending the relative reward strength algorithm proposed by Simha and Kurose. The learning property of our algorithm is considered theoretically, and it is proved that the optimal path probability can be approached to 1 as close as possible by using our algorithm. In numerical simulation, the number of iterations of our algorithm is compared with that of the variable hierarchical structure learning algorithm of LR-I type proposed by Mogami and Baba, and it is shown that our algorithm can find the optimal path after the smaller number of iterations than that of the algorithm of LR-I type.
