Traffic-sensitive controllers for multi-car elevators; design, multi-objective optimization and analysis

Multi-car elevator (MCE) that has several elevator cars in a single shaft is the novel system and attracts attention for improvement of transportation in high-rise buildings. Design of controller for MCE is very difficult engineering problem, and one of the promising approaches is from-scratch evolutionary optimization of the controller through discrete event simulation of the MCE system. In the present paper, the authors at first show application of evolutionary multi-objective optimization to design traffic-sensitive MCE controller. The controller for MCE is optimized for different traffic conditions by combining the multi-objective optimization with the exemplar-based policy (EBP) representation that has adequate flexibility and generalization ability as a controller. And a policy visualization method is proposed for analyzing how the traffic-sensitive EBP works.