An online tuning method for multiobjective control of elevator group

The authors describe an online tuning method of control parameters for a multiobjective elevator group control system. To adapt a control parameter to each building, a conventional system tunes the control parameter by using a combined method of a learning function and a built-in simulator. However, an increase of tuning time, caused by the addition of control parameters, can become a serious problem. The proposed method can restrict research points by using macro knowledge of the system domain. Then, tuning time and tuning accuracy are well-balanced. Computer simulations show quick settling and good tuning results. Field data obtained with the system verify the findings