基于改进深度极限学习机的多热点区域出租车需求预测

doi:10.3969/j.issn.1674-0696.2026.06.09

摘要/Abstract

摘要： 针对城市出租车的时空需求分布不均衡、多热点区域时变性强的问题，提出基于改进深度极限学习机的多热点区域出租车需求预测方法。首先，通过变分模态分解方法将原始需求序列分解为多组具有明确物理意义的模态分量，有效分离不同时间尺度的需求波动特征；然后，在构建深度极限学习机模型的基础上利用基于种群的元启发式优化算法优化深度极限学习机的权重与偏差参数；最后，通过纽约曼哈顿6个热点区域的出租车需求数据验证该方法的有效性。结果表明：改进深度极限学习机能够在不同热点区域均实现精准的出租车需求预测；与其他基线模型和单一模型相比，改进深度极限学习机订单量平均绝对误差EMA、均方根误差ERMS和平均绝对百分比误差EMAP分别降低了2.43~8.54个/h、3.01~11.77个/h、0.54%~2.91%。

关键词: 交通运输工程；出租车需求预测；变分模态分解算法；深度极限学习机；混沌进化优化算法

Abstract: Aiming at the uneven distribution of spatiotemporal demand for urban taxis and the strong temporal variability in multiple hotspot areas, a taxi demand prediction method based on improved deep extreme learning machine (DELM) in multiple hotspot areas was proposed. Firstly, the variational mode decomposition method was applied to decompose the original demand series into multiple sets of modal components with clear physical meanings, effectively separating demand fluctuation features at different time scales. Then, based on the construction of the deep extreme learning machine model, a population-based metaheuristic optimization algorithm was employed to optimize the weights and bias parameters of the DELM. Finally, the effectiveness of the proposed method was validated through taxi demand data from six hotspots in Manhattan, New York. The results demonstrate that the improved DELM achieves accurate taxi demand predictions across different hotspot areas, with statistically significant reductions in EMA, ERMS and EMAP of order volume by 2.43~8.54 orders per hour, 3.01~11.77 orders per hour, and 0.54%~2.91%, respectively, compared to other baseline models and single models.

Key words: traffic and transportation engineering; taxi demand forecasting; variational mode decomposition algorithm; deep extreme learning machine; chaotic evolutionary optimization algorithm

中图分类号:

U492

车畅畅，李浩. 基于改进深度极限学习机的多热点区域出租车需求预测[J]. 重庆交通大学学报（自然科学版）, 2026, 45(6): 64-75.

CHE Changchang, LI Hao. Taxi Demand Prediction in Multiple Hotspot Areas Based on Improved Deep Extreme Learning Machine[J]. Journal of Chongqing Jiaotong University(Natural Science), 2026, 45(6): 64-75.

/ 推荐

参考文献

［1］王君悦,董春娇,王菁,等.轨迹数据驱动的出租车出行特征分析与时空分布建模［J/OL］.吉林大学学报(工学版),2025-05-22.https:∥doi.org/10.13229/j.cnki.jdxbgxb.20250061.
WANG Junyue，DONG Chunjiao，WANG Jing，et al. Trajectory data-driven taxi travel characteristics analysis and spatial-temporal distribution modeling ［J/OL］. Journal of Jilin University (Engineering and Technology Edition), 2025-05-22. https:∥doi.org/10.13229/j.cnki.jdxbgxb.20250061.
［2］ YANGYuwei, LI Zhuoxuan, CHEN Jun, et al. TRELM-DROP: an impavement non-iterative algorithm for traffic flow forecast［J］. Physica A: Statistical Mechanics and Its Applications, 2024, 633: 129337.
［3］马健霄, 印杰祥, 邵春福, 等. 时空异质性建成环境对出租车出行需求影响［J］. 北京交通大学学报, 2025, 49(1): 119-127.
MA Jianxiao, YIN Jiexiang, SHAO Chunfu, et al. Impact of spatiotemporal heterogeneity in built environment on taxi travel demand［J］. Journal of Beijing Jiaotong University, 2025, 49(1): 119-127.
［4］ LIM B, ZOHREN S. Time-series forecasting with deep learning: a survey［J］. Philosophical Transactions: Mathematical, Physical and Engineering Sciences, 379(2194): 20200209.
［5］ HUGuojing, WHALIN R W, KWEMBE T A, et al. Short-term traffic flow prediction based on secondary hybrid decomposition and deep echo state networks［J］. Physica A: Statistical Mechanics and its Applications, 2023, 632: 129313.
［6］程静,刘家骏,高勇.基于时间序列聚类方法分析北京出租车出行量的时空特征［J］.地球信息科学学报,2016,18(9):1227-1239.
CHENG Jing, LIU Jiajun, GAO Yong. Analyzing the spatio-temporal characteristics of Beijing’s OD trip volume based on time series clustering method ［J］. Journal of Geo-information Science, 2016, 18(9): 1227-1239.
［7］王晓全,邵春福,尹超英,等.基于ARIMA-GARCH-M模型的短时交通流预测方法［J］.北京交通大学学报,2018,42(4):79-84.
WANG Xiaoquan, SHAO Chunfu, YIN Chaoying, et al. Short term traffic flow forecasting method based on ARIMA-GARCH-M model ［J］. Journal of Beijing Jiaotong University, 2018, 42(4): 79-84.
［8］ DAVIS N, RAINA G, JAGANNATHAN K. A multi-level clustering approach for forecasting taxi travel demand［C］∥2016 IEEE 19th International Conference on Intelligent Transportation Systems (ITSC). ACM, 2016: 223-228.
［9］ KE Jintao, ZHENG Hongyu, YANG Hai, et al. Short-term forecasting of passenger demand under on-demand ride services: a spatio-temporal deep learning approach［J］. Transportation Research Part C: Emerging Technologies, 2017, 85: 591-608.
［10］ FAGHIH S, SHAH A, WANGZehan, et al. Taxi and mobility: modeling taxi demand using ARMA and linear regression［J］. Procedia Computer Science, 2020, 177: 186-195.
［11］江建霖.基于组合模型的机场客流量预测方法研究［J］.长江信息通信,2024,37(8):88-92.
JIANG Jianlin. Research on airport passenger flow prediction method based on combined model［J］. Changjiang Information & Communications, 2024, 37(8): 88-92.
［12］李鹏程.基于机器学习方法的出租车出行乘客需求预测［D］. 上海：上海交通大学, 2019.
LI Pengcheng. The Prediction of Taxi Demand Based on Machine Learning ［D］. Shanghai：Shanghai Jiao Tong University, 2019. 001430.
［13］ LIUZhizhen, CHEN Hong, SUN Xiaoke, et al. Data-driven real-time online taxi-hailing demand forecasting based on machine learning method［J］. Applied Sciences, 2020, 10(19): 6681.
［14］李岩.基于热点区域的出租车出行需求预测模型研究［D］.西安：长安大学,2019.
LI Yan. Research on Taxi Travel Demand Forecasting Model Based on Hot Areas ［D］. Xi’an：Chang’an University, 2019.
［15］丁新宇,施佺.基于深度时空残差网络的路网短时交通流预测［J］.计算机工程与设计,2022,43(2):316-322.
DING Xinyu, SHI Quan. Short-term traffic flow prediction of road network based on deep spatio-temporal residual network［J］. Computer Engineering and Design, 2022, 43(2): 316-322.
［16］段宗涛, 张凯, 杨云, 等. 基于深度CNN-LSTM-ResNet组合模型的出租车需求预测［J］. 交通运输系统工程与信息, 2018, 18(4): 215-223.
DUAN Zongtao, ZHANG Kai, YANG Yun, et al. Taxi demand prediction based on CNN-LSTM-ResNet hybrid depth learning model［J］. Journal of Transportation Systems Engineering and Information Technology, 2018, 18(4): 215-223.
［17］ LIAO Wenxiong, ZENG Bi, LIU Jianqi, et al. Taxi demand forecasting based on the temporal multimodal information fusion graph neural network［J］. Applied Intelligence, 2022, 52(10): 12077-12090.
［18］ KE Jintao, QIN Xiaoran, YANG Hai, et al. Predicting origin-destination ride-sourcing demand with a spatio-temporal encoder-decoder residual multi-graph convolutional network［J］. Transportation Research Part C: Emerging Technologies, 2021, 122: 102858.
［19］ GENG Xu, LI Yaguang, WANG Leye, et al. Spatiotemporal multi-graph convolution network for ride-hailing demand forecasting［J］. Proceedings of the AAAI Conference on Artificial Intelligence, 2019, 33(1): 3656-3663.
［20］ XU Jun, RAHMATIZADEH R, BLNI L, et al. Real-time prediction of taxi demand using recurrent neural networks［J］. IEEE Transactions on Intelligent Transportation Systems, 2018, 19(8): 2572-2581.
［21］ ALAM S, AYUB M S, CUIHao, et al. A comparative study of machine learning models for taxi-demand prediction using a big data framework［J］. Public Transport, 2025, 17(3): 803-833.
［22］段锦, 姚安妮, 王震, 等. 改进的麻雀搜索算法优化无线传感器网络覆盖［J］. 吉林大学学报(工学版), 2024, 54(3): 761-770.
DUAN Jin, YAO Anni, WANG Zhen, et al. Improved sparrow search algorithm optimizes coverage in wireless sensor networks［J］. Journal of Jilin University (Engineering and Technology Edition), 2024, 54(3): 761-770.
［23］ DONG Yingchao, ZHANG Shaohua, ZHANG Hongli, et al. Chaotic evolution optimization: a novel metaheuristic algorithm inspired by chaotic dynamics［J］. Chaos, Solitons & Fractals, 2025, 192: 116049.
［24］康辉，段寒，陶依，等.基于切比雪夫对立的二进制蝙蝠算法的特征选择多目标优化的生物启发式方法［J/OL］.吉林大学学报(工学版)，2025-01-24.https:∥doi.org/10.13229/j.cnki.jdxbgxb.20241220.
KANG Hui, DUAN Han, TAO Yi, et al. Bio-Inspired approach for multi-objective optimization of feature selection using chebyshev opposite-based binary bat algorithm ［J/OL］. Journal of Jilin University (Engineering and Technology Edition)，2025-01-24. https:∥doi.org/10.13229/j.cnki.jdxbgxb.20241220.
［25］ TIZHOOSH H R. Opposition-based learning: a new scheme for machine intelligence［C］∥ International Conference on Computational Intelligence for Modelling, Control and Automation and International Conference on Intelligent Agents, Web Technologies and Internet Commerce (CIMCA-IAWTIC’06). IEEE, 2006.
［26］王改革, 郭立红, 段红, 等. 基于萤火虫算法优化BP神经网络的目标威胁估计［J］. 吉林大学学报(工学版), 2013, 43(4): 1064-1069.
WANG Gaige, GUO Lihong, DUAN Hong, et al. Target threat assessment using glowworm swarm optimization and BP neural network［J］. Journal of Jilin University (Engineering and Technology Edition), 2013, 43(4): 1064-1069.
［27］邹红波, 李奇隆. 基于改进北方苍鹰优化算法的激光焊缝图像分类［J］. 吉林大学学报(工学版), 2025, 55(4): 1426-1435.
ZOU Hongbo, LI Qilong. Laser weld image classification based on improved northern goshawk optimization algorithm［J］. Journal of Jilin University (Engineering and Technology Edition), 2025, 55(4): 1426-1435.

[1]	周旦1，孙家煜2，顾国斌1，钟楚捷3，王涛1. 基于轨迹数据的出租车司机寻客路径优化方法[J]. 重庆交通大学学报（自然科学版）, 2024, 43(1): 83-90.
[2]	朱顺应，李维吉，林小媛，肖文彬，王红. 特色小镇游客车辆无复检管控点布设研究[J]. 重庆交通大学学报（自然科学版）, 2021, 40(06): 52-58.
[3]	朱顺应1，刘滨1，刘斌2，杨友珍3，王红1. 考虑运营收益与车厢空间感知的轨道列车发车间隔优化[J]. 重庆交通大学学报（自然科学版）, 2021, 40(04): 41-47.
[4]	陶健，王文君，庞夺峰，殷传峰. 高速公路危险货物运输人员暴露风险实时评价分析[J]. 重庆交通大学学报（自然科学版）, 2021, 40(03): 43-48.
[5]	计斌，陈瑶，王璐，李昊明，常国威. 城市客运出租汽车运力投放社会稳定风险评估模型[J]. 重庆交通大学学报（自然科学版）, 2020, 39(12): 26-30.
[6]	熊杰1，白宸宇1，李向楠2，陈艳艳1. 整车充电模式下的纯电动公交行车计划优化[J]. 重庆交通大学学报（自然科学版）, 2020, 39(11): 38-44.
[7]	牛世峰1，李贵强2，刘恩全3. 基于Logit模型的城乡结合部不同车型安全性分析[J]. 重庆交通大学学报（自然科学版）, 2020, 39(01): 28-33.
[8]	彭辉，赵亚军. 基于DEA模型的城市公交交通效益对比评价[J]. 重庆交通大学学报（自然科学版）, 2019, 38(12): 110-114.
[9]	徐婷，姜瑞森，李洪庆，李青，陈刚. 矩形网格道路多模式公交线网布局优化研究[J]. 重庆交通大学学报（自然科学版）, 2019, 38(10): 108-115.
[10]	李绍春1，初永玲1，廉静2，廖宝梁3，吕承举3，纪少波2. BP神经网络在液化天然气客车能耗估算中的应用[J]. 重庆交通大学学报（自然科学版）, 2019, 38(10): 133-138.
[11]	陈立辉1,2 ，郭忠印1. 高速公路连续下坡路段大货车速度特征研究[J]. 重庆交通大学学报（自然科学版）, 2019, 38(08): 86-91.
[12]	王征，陶健，王文君. 城市快速路危险货物运输风险动态评估[J]. 重庆交通大学学报（自然科学版）, 2019, 38(06): 96-102.
[13]	田丽君，詹晓斌. 考虑交通拥堵的出租车市场均衡模型[J]. 重庆交通大学学报（自然科学版）, 2019, 38(01): 76-82.
[14]	陈艳艳1，秦伟玲1，李晓祎2，吴克寒1，陈宁1，宋程程1. 货运场站机动车污染物排放监测及污染特征研究[J]. 重庆交通大学学报（自然科学版）, 2018, 37(09): 60-65.
[15]	靳文舟，叶钦海，郝小妮. 基于用户预约数据的公共自行车高峰期调度研究[J]. 重庆交通大学学报（自然科学版）, 2017, 36(10): 91-96.