Shintaro Tsuji

Application of the expert system to elevator group-supervisory control

A novel elevator group-supervisory control method to which an expert system and fuzzy rules are applied, is suggested. This method determines the optimum car to answer a hall call by using knowledge bases, production rules and fuzzy rules, extracted from the knowledge of experts, in elevator group-supervisory control. It has been verified through simulation that the method can bring about considerable improvement, reducing the average waiting time by 15-20%, and long waits (not less than 60s) 30-50% in comparison with the conventional method, in which a fixed evaluation function is used. Knowledge-based systems suitable to the online and real-time control for the implementation of this method on 32-b microcomputers are suggested. The result of the simulation has demonstrated that the suggested knowledge-based systems satisfy the required response time of 100-150 ms in the group-supervisory control system with an immediate prediction function.<<ETX>>