The neural network imitates the information processing technology which a brain has. This system can perform the pattern recognition and the association memory like human brains do.
In this research, a neural network with internal memory proceeded with a simulation in the game. This is a game in which a player captures some targets in a square field.
Using this game, we checked whether the control would be possible when the supervising signal in consideration of change in time is used, and we studied about the influence on the neural network.
As a result, this time series control was successful.
However, in the case with these supervising signals, tha gap between the neural network output and the supervising signal became large, and efficiency became worse, compared with the case when the supervising signal out of consideration of change in time is used.
It is a future subject to improve the supervising signal in consideration of change in time, and to investigate the influence on the neural network on this supervising signal further.