组合模型在电梯客流量预测中的仿真研究

摘要

Research elevator traffic flow prediction problems. The elevator flow has the characteristics of cyclical change, randomness and nonlinear, traditional prediction method is based on linear data predictive model and cannot accurately describe the nonlinear elevator traffic, and the prediction accuracy is low. In order to improve the prediction precision, this paper proposed a combination prediction model based on ARM A and RBF neural network. Firstly, the linear part of televator traffic flow was predicted by AR1MA, then the non-linear part was predicted by RBF neural network, finally combination model prediction results were obtained. Simulation results indicate that the combined model has the advantages of ARMA and RBF neural network, and improves the prediction accuracy. It provides a new analytical tools for elevator traffic flow prediction.%研究电梯客流量准确预测问题,以保证电梯运行安全.电梯客流量受到周末、上下班及假期影响,流量具有周期性、随机性和非线性变化特点.传统预测模型难以准确描述动态特点,导致电梯客流量的预测准确率低.为了提高电梯客流量的预测准确率,提出一种ARMA和RBF神经网络相结合的电梯客流量组合预测模型.组合模型首先利用ARIMA对电梯客流量线性变化部分进行预测,然后采用RBF神经网络对非线性部分进行预测,最后将两者结果相加,利用组合模型进行电梯客流量预测.仿真结果表明,组合模型用ARMA和RBF神经网络的优点,提高了电梯客流量的预测准确率,为电梯调度及优化控制提供了一种新的分析方法.