In this paper, we propose an efficient control method for multicar-elevator using reinforcement learning．In the method, the control agent selects the best policy among three policies, namely Distance-policy, MinGuests-policy, and Zone-policy, according to traffic flow. The control agent takes the number of total passengers and the distance from the starting floor to the destination floor of a call into account. Through experiments, we compare the mean completion time of the proposed method with the mean completion times for the cases when the car assignment is made by each of the three policies.

2011 IEEE SMC Hiroshima Chapter若手研究会, 2011年7月9日, 広島市立大学