抄録
In this paper, we introduce a method of space division called adaptive state spaces division with N-actions during the learning process. In the proposed method, we arrange the temporary mother points into lattice structure firstly. When we act and learn by temporary mother points, the new mother points are generated in the state space with the number of actions by the accumulation data of the state transition vector. We delete the temporary mother points in which the new points are added, and divide the state space into a number of actions subspaces.
