基于多Agent的电梯智能群控系统

给出了基于Agent的多电梯智能群控系统结构,并对其中的状态协调Agent以及调度Agent进行了深入研究,在分析电梯乘客需求的随机性与周期性特点的基础上,给出了建立客流模板的学习算法.根据乘客的现实需求和预测得出的潜在需求对梯群预测调度,给出了选择客流模板的算法和基于乘客需求预测的最大引力派梯算法.该方法能够利用预测的乘客需求,尽量减小调度的滞后性,从而减少乘客的平均候梯时间及长时间等待率.仿真结果显示了该方法的优越性.

Elevator intelligent group control system based on multi-Agent

On the basis of multi-Agent technology,a structure of multi elevator group control system(EGCS) is put forward and the state harmonizing Agent and dispatching Agent in EGCS are investigated.By analyzing the randomicity and periodicity of the passengers demand,the study arithmetic setting up the passenger flow template is applied.The forecast and dispatch of multi elevator group is brought forward by the potential demand of the passenger's current demand and prediction.After the choosing passengers template arithmetic is applied, the dispatch template is advanced,which can reduce the dispatch lag,as well as the passenger waiting time.Simulation analysis results demonstrate the validity of this control method.