Stochastic Control for Multi Car Elevator by using Learning Automatons

Abstract

An adaptive control technique for multi car elevators (MCEs) with learning automatons: LAs is proposed. MCEs are high performance, near-future elevator systems with multi shafts and multi cars. A strong point of the novel elevators is that realizing a large carrying capacity in small shaft area. For the purpose of realizing the elevators, adapting to dynamic environment is a significant issue in the real world application. In order to realize autonomous behavior as the simple traffic control, LA is one of the appropriate approach. Where, LA has a stochastic automaton as each car control system. Each stochastic automaton varies its stochastic behavior distributions for adapting to environment in which the policy is evaluated with each passenger waiting times. Accordingly, LA learns the environment autonomously. By using the LA as the basis of control technique, the efficiency of MCE control is evaluated through simulation experiments. The results show the technique enables us to reduce waiting times and to improve efficiently.