基于SSA-LSTM组合模型的城市轨道交通短时客流预测  

Short-term prediction of urban railtransit passenger flow based on the Sparrow Search Algorithm-Long Short Term Memory combination model

作　　者：姜嘉伟 赵金宝[1,2] 刘文静 徐月娟 李明星 JIANG Jiawei;ZHAO Jinbao;LIU Wenjing;XU Yuejuan;LI Mingxing(School of Transportation and Vehicle Engineering,Shandong University of Technology,Zibo 255000,China;School of Transportation,Southeast University,Nanjing 210009,China)

机构地区：[1]山东理工大学交通与车辆工程学院,山东淄博255000 [2]东南大学交通学院,江苏南京210009

出　　处：《山东科学》2023年第5期75-84,共10页Shandong Science

基　　金：国家自然科学基金项目(51608313);山东省自然科学基金项目(ZR2021MF109);山东高速集团科技项目(2020-SDHS-GSJT-024)。

摘　　要：随着我国经济的快速增长及城市化水平的不断提高,轨道交通在居民出行中发挥着越来越重要的作用。作为影响城市轨道交通运营效益和服务水平的关键因素,客流精准预测受到运营管理者和研究者的日益重视。为提高城市轨道交通客流预测精度,提出了基于麻雀搜索算法(SSA)和长短期记忆网络(LSTM)的SSA-LSTM组合模型。本文以杭州地铁一号线客流量数据为例,在选取轨道交通客流相关影响因素的基础上,利用建立的SSA-LSTM模型对相关站点进行短时客流预测,并与LSTM模型、遗传算法(GA)优化的LSTM模型(GA-LSTM)以及粒子群算法(PSO)优化的LSTM模型(PSO-LSTM)预测结果进行对比分析。结果表明,相比于前述参照模型,SSA-LSTM模型的预测精度分别提升了19.1%、9.7%和2.4%,并在均方根误差指标方面有更优异的表现。SSA-LSTM组合模型在城市轨道交通客流预测中具有一定的应用价值,具有协助运营管理者提高城市轨道交通运营管理效益和提高服务水平的潜力。With the rapid growth of China′s economy and the continuous urbanization,rail transit plays an increasingly important role in residents′travel.As a key factor affecting the operation efficiency and service level of urban rail transit,accurate passenger flow prediction has attracted increasing attention from operation managers and researchers.To improve the prediction accuracy of the urban rail transit passenger flow,this paper combines sparrow search algorithm(SSA)and long short-term memory network(LSTM)and proposed a SSA-LSTM combined model.Based on the passenger flow data obtained from four stations of Hangzhou Metro Line 1 and the selected factors affecting the rail transit passenger flow,we used the proposed SSA-LSTM model to predict the short-term passenger flow of relevant stations.Then,we compared the predicted results with those estimated by the LSTM,GA-LSTM,and PSO-LSTM models.Results show that the prediction accuracy of the proposed model is 16.0%,8.8%,and 2.3%,higher than the aforementioned models,respectively;furthermore,the proposed method exhibited better performance in terms of the root mean square error.Thus,the proposed model has potential applicationin predicting the urban rail transit passenger flow.Moreover,it can assistoperation managers in improving the operation efficiency and service level of urban rail transit.

关 键 词：城市轨道交通 短时客流预测 麻雀搜索算法 长短期记忆网络 组合模型 

分 类 号：U231[交通运输工程—道路与铁道工程]