Predicting citywide passenger demand via reinforcement learning from spatio-temporal dynamics

Xiaodong Ning, Lina Yao, Xianzhi Wanga, Boualem Benatallah , Flora Salim, Pari Delir Haghighi

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearchpeer-review

Abstract

The global urbanization imposes unprecedented pressure on urban infrastructure and public resources. The population explosion has made it challenging to satisfy the daily needs of urban residents. 'Smart City' is a solution that utilizes different types of data collection sensors to help manage assets and resources intelligently and more efficiently. Under the Smart City umbrella, the primary research initiative in improving the efficiency of car-hailing services is to predict the citywide passenger demand to address the imbalance between the demand and supply. However, predicting the passenger demand requires analysis on various data such as historical passenger demand, crowd outflow, and weather information, and it remains challenging to discover the latent relationships among these data. To address this challenge, we propose to improve the passenger demand prediction via learning the salient spatial-temporal dynamics within a reinforcement learning framework. Our model employs an information selection mechanism to focus on the most distinctive data in historical observations. This mechanism can automatically adjust the information zone according to the prediction performance to find the optimal choice. It also ensures the prediction model to take full advantage of the available data by introducing the positive and excluding the negative correlations. We have conducted experiments on a large-scale real-world dataset that covers 1.5 million people in a major city in China. The results show our model outperforms state-of-the-art and a series of baselines by a large margin.
Original languageEnglish
Title of host publicationProceedings of the 15th EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services
Subtitle of host publication5-7 November 2018, New York City, NY, United States - Mobiquitous 2018
EditorsCristian Borcea, Shiwen Mao, Jian Tang
Place of PublicationNew York NY USA
PublisherAssociation for Computing Machinery (ACM)
Number of pages10
ISBN (Electronic)9781450360937
DOIs
Publication statusPublished - 2018
EventInternational Conference on Mobile and Ubiquitous Systems: Networks and Services 2018 - New York, United States of America
Duration: 5 Nov 20187 Nov 2018
Conference number: 15th
http://mobiquitous2018.eai-conferences.org/

Conference

ConferenceInternational Conference on Mobile and Ubiquitous Systems: Networks and Services 2018
Abbreviated titleMobiQuitous 2018
CountryUnited States of America
CityNew York
Period5/11/187/11/18
Internet address

Cite this

Ning, X., Yao, L., Wanga, X., Benatallah , B., Salim, F., & Delir Haghighi, P. (2018). Predicting citywide passenger demand via reinforcement learning from spatio-temporal dynamics. In C. Borcea, S. Mao, & J. Tang (Eds.), Proceedings of the 15th EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services: 5-7 November 2018, New York City, NY, United States - Mobiquitous 2018 New York NY USA: Association for Computing Machinery (ACM). https://doi.org/10.1145/3286978.3286991
Ning, Xiaodong ; Yao, Lina ; Wanga, Xianzhi ; Benatallah , Boualem ; Salim, Flora ; Delir Haghighi, Pari. / Predicting citywide passenger demand via reinforcement learning from spatio-temporal dynamics. Proceedings of the 15th EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services: 5-7 November 2018, New York City, NY, United States - Mobiquitous 2018. editor / Cristian Borcea ; Shiwen Mao ; Jian Tang. New York NY USA : Association for Computing Machinery (ACM), 2018.
@inproceedings{6a62b95f35864263af363e6540292396,
title = "Predicting citywide passenger demand via reinforcement learning from spatio-temporal dynamics",
abstract = "The global urbanization imposes unprecedented pressure on urban infrastructure and public resources. The population explosion has made it challenging to satisfy the daily needs of urban residents. 'Smart City' is a solution that utilizes different types of data collection sensors to help manage assets and resources intelligently and more efficiently. Under the Smart City umbrella, the primary research initiative in improving the efficiency of car-hailing services is to predict the citywide passenger demand to address the imbalance between the demand and supply. However, predicting the passenger demand requires analysis on various data such as historical passenger demand, crowd outflow, and weather information, and it remains challenging to discover the latent relationships among these data. To address this challenge, we propose to improve the passenger demand prediction via learning the salient spatial-temporal dynamics within a reinforcement learning framework. Our model employs an information selection mechanism to focus on the most distinctive data in historical observations. This mechanism can automatically adjust the information zone according to the prediction performance to find the optimal choice. It also ensures the prediction model to take full advantage of the available data by introducing the positive and excluding the negative correlations. We have conducted experiments on a large-scale real-world dataset that covers 1.5 million people in a major city in China. The results show our model outperforms state-of-the-art and a series of baselines by a large margin.",
author = "Xiaodong Ning and Lina Yao and Xianzhi Wanga and Boualem Benatallah and Flora Salim and {Delir Haghighi}, Pari",
year = "2018",
doi = "10.1145/3286978.3286991",
language = "English",
editor = "Cristian Borcea and Mao, {Shiwen } and Jian Tang",
booktitle = "Proceedings of the 15th EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services",
publisher = "Association for Computing Machinery (ACM)",
address = "United States of America",

}

Ning, X, Yao, L, Wanga, X, Benatallah , B, Salim, F & Delir Haghighi, P 2018, Predicting citywide passenger demand via reinforcement learning from spatio-temporal dynamics. in C Borcea, S Mao & J Tang (eds), Proceedings of the 15th EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services: 5-7 November 2018, New York City, NY, United States - Mobiquitous 2018. Association for Computing Machinery (ACM), New York NY USA, International Conference on Mobile and Ubiquitous Systems: Networks and Services 2018, New York, United States of America, 5/11/18. https://doi.org/10.1145/3286978.3286991

Predicting citywide passenger demand via reinforcement learning from spatio-temporal dynamics. / Ning, Xiaodong ; Yao, Lina; Wanga, Xianzhi; Benatallah , Boualem ; Salim, Flora; Delir Haghighi, Pari.

Proceedings of the 15th EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services: 5-7 November 2018, New York City, NY, United States - Mobiquitous 2018. ed. / Cristian Borcea; Shiwen Mao; Jian Tang. New York NY USA : Association for Computing Machinery (ACM), 2018.

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearchpeer-review

TY - GEN

T1 - Predicting citywide passenger demand via reinforcement learning from spatio-temporal dynamics

AU - Ning, Xiaodong

AU - Yao, Lina

AU - Wanga, Xianzhi

AU - Benatallah , Boualem

AU - Salim, Flora

AU - Delir Haghighi, Pari

PY - 2018

Y1 - 2018

N2 - The global urbanization imposes unprecedented pressure on urban infrastructure and public resources. The population explosion has made it challenging to satisfy the daily needs of urban residents. 'Smart City' is a solution that utilizes different types of data collection sensors to help manage assets and resources intelligently and more efficiently. Under the Smart City umbrella, the primary research initiative in improving the efficiency of car-hailing services is to predict the citywide passenger demand to address the imbalance between the demand and supply. However, predicting the passenger demand requires analysis on various data such as historical passenger demand, crowd outflow, and weather information, and it remains challenging to discover the latent relationships among these data. To address this challenge, we propose to improve the passenger demand prediction via learning the salient spatial-temporal dynamics within a reinforcement learning framework. Our model employs an information selection mechanism to focus on the most distinctive data in historical observations. This mechanism can automatically adjust the information zone according to the prediction performance to find the optimal choice. It also ensures the prediction model to take full advantage of the available data by introducing the positive and excluding the negative correlations. We have conducted experiments on a large-scale real-world dataset that covers 1.5 million people in a major city in China. The results show our model outperforms state-of-the-art and a series of baselines by a large margin.

AB - The global urbanization imposes unprecedented pressure on urban infrastructure and public resources. The population explosion has made it challenging to satisfy the daily needs of urban residents. 'Smart City' is a solution that utilizes different types of data collection sensors to help manage assets and resources intelligently and more efficiently. Under the Smart City umbrella, the primary research initiative in improving the efficiency of car-hailing services is to predict the citywide passenger demand to address the imbalance between the demand and supply. However, predicting the passenger demand requires analysis on various data such as historical passenger demand, crowd outflow, and weather information, and it remains challenging to discover the latent relationships among these data. To address this challenge, we propose to improve the passenger demand prediction via learning the salient spatial-temporal dynamics within a reinforcement learning framework. Our model employs an information selection mechanism to focus on the most distinctive data in historical observations. This mechanism can automatically adjust the information zone according to the prediction performance to find the optimal choice. It also ensures the prediction model to take full advantage of the available data by introducing the positive and excluding the negative correlations. We have conducted experiments on a large-scale real-world dataset that covers 1.5 million people in a major city in China. The results show our model outperforms state-of-the-art and a series of baselines by a large margin.

U2 - 10.1145/3286978.3286991

DO - 10.1145/3286978.3286991

M3 - Conference Paper

BT - Proceedings of the 15th EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services

A2 - Borcea, Cristian

A2 - Mao, Shiwen

A2 - Tang, Jian

PB - Association for Computing Machinery (ACM)

CY - New York NY USA

ER -

Ning X, Yao L, Wanga X, Benatallah B, Salim F, Delir Haghighi P. Predicting citywide passenger demand via reinforcement learning from spatio-temporal dynamics. In Borcea C, Mao S, Tang J, editors, Proceedings of the 15th EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services: 5-7 November 2018, New York City, NY, United States - Mobiquitous 2018. New York NY USA: Association for Computing Machinery (ACM). 2018 https://doi.org/10.1145/3286978.3286991