Scheduling Of Multi-Objective Elevator Group Control Algorithm Based On AIS

For the past few years, with the fast development of modern city and high-risebuildings, higher service quality of elevator is needed. Lots of advanced group controlalgorithms have been widely applied in kinds of elevator group control systems, andthe performance of elevator group control system has been further developed as well.However, many problems still stand ahead and should be solved urgently.Because the traditional elevator which has only two buttons in the elevator lobby,has made the calling elevator signal and destination signal separate, when we arewaiting for the elevator, only passengers’ up or down information of elevator groupcontrol system can be acquired, and the idiographic destination information is not begot until passengers get into the elevator, which spells that it’s difficult to predictpassengers’ waiting time and riding time. Although putting the selecting floor panelinto the elevator lobby, elevator group control system with destination registration cansolve separation signal problem in advance, it can not calculate the number of thesame starting-arriving floor passengers, which also can cause some passengers can’tget into the elevator with others at a time, influence the forecast of waiting time andwaste unnecessary elevator resource.To avoid a serial of problems existed in traditional elevator group control system,such as incompleteness of passengers’ information, large energy consumption ofelevator group and the difficulty to guarantee service quality, this paper proposed anew elevator group control system with destination registration which is based on theprimary elevator group control system with destination registration. We let touch-typeintelligent LCD take place of the traditional lobby calling panel, which likes “BankQueue System”. Using intelligent LCD elevator group control system will allocatenumber for every passenger, retrace a corresponding riding time program by systemcalculating. According to the acquired information, the new elevator group controlsystem with destination registration can more accurately calculate the passengers’ riding time, waiting time and degree of elevator congestion, reduce the passengers’long waiting proportion, and get a more reasonable dispatching program. Taking intoaccount passengers’ riding time, waiting time, high waiting proportion, elevator energyconsumption and degree of elevator congestion, I proposed artificial immunealgorithm which combined with gradient descent algorithm according to theadvantages of each other. As collecting all passengers flow information, the elevatorcan visually judge the different traffic modes, and achieve the optimal elevator groupcontrol under different conditions.At last, I compiled simulation program of elevator group control system withsoftware of Matlab. The result showed that, under different traffic modes, artificialimmune algorithm which combined with gradient descent algorithm has much moreobvious validity and superiority than the standard artificial immune algorithm.