Abstract
An application of variable-structure stochastic automata to commercial games is considered. It is shown that the utilization of the learning performances of stochastic automata is quite effective for commercial games with incomplete information.