Analyzing the Impact of Traffic Congestion Mitigation: From an Explainable Neural Network Learning Framework to Marginal Effect Analyses

Jianping Sun, Jifu Guo, Xin Wu, Qian Zhu, Danting Wu, Kai Xian, Xuesong Zhou

Research output: Contribution to journalArticle

Abstract

Computational graphs (CGs) have been widely utilized in numerical analysis and deep learning to represent directed forward networks of data flows between operations. This paper aims to develop an explainable learning framework that can fully integrate three major steps of decision support: Synthesis of diverse traffic data, multilayered traffic demand estimation, and marginal effect analyses for transport policies. Following the big data-driven transportation computational graph (BTCG) framework, which is an emerging framework for explainable neural networks, we map different external traffic measurements collected from household survey data, mobile phone data, floating car data, and sensor networks to multilayered demand variables in a CG. Furthermore, we extend the CG-based framework by mapping different congestion mitigation strategies to CG layers individually or in combination, allowing the marginal effects and potential migration magnitudes of the strategies to be reliably quantified. Using the TensorFlow architecture, we evaluate our framework on the Sioux Falls network and present a large-scale case study based on a subnetwork of Beijing using a data set from the metropolitan planning organization.

Original languageEnglish (US)
JournalSensors (Basel, Switzerland)
Volume19
Issue number10
DOIs
StatePublished - May 15 2019
Externally publishedYes

Fingerprint

congestion
Traffic congestion
Mobile phones
learning
traffic
Sensor networks
Numerical analysis
Railroad cars
Learning
Neural networks
Planning
Cell Phones
Policy Making
Organizations
floating
numerical analysis
planning
emerging
Big data
Deep learning

Keywords

  • computational graph
  • congestion mitigation
  • marginal analyses
  • TensorFlow
  • traffic demand estimation

ASJC Scopus subject areas

  • Analytical Chemistry
  • Atomic and Molecular Physics, and Optics
  • Biochemistry
  • Instrumentation
  • Electrical and Electronic Engineering

Cite this

Analyzing the Impact of Traffic Congestion Mitigation : From an Explainable Neural Network Learning Framework to Marginal Effect Analyses. / Sun, Jianping; Guo, Jifu; Wu, Xin; Zhu, Qian; Wu, Danting; Xian, Kai; Zhou, Xuesong.

In: Sensors (Basel, Switzerland), Vol. 19, No. 10, 15.05.2019.

Research output: Contribution to journalArticle

@article{96c80faf77604bd7a6b03e9b72513ca4,
title = "Analyzing the Impact of Traffic Congestion Mitigation: From an Explainable Neural Network Learning Framework to Marginal Effect Analyses",
abstract = "Computational graphs (CGs) have been widely utilized in numerical analysis and deep learning to represent directed forward networks of data flows between operations. This paper aims to develop an explainable learning framework that can fully integrate three major steps of decision support: Synthesis of diverse traffic data, multilayered traffic demand estimation, and marginal effect analyses for transport policies. Following the big data-driven transportation computational graph (BTCG) framework, which is an emerging framework for explainable neural networks, we map different external traffic measurements collected from household survey data, mobile phone data, floating car data, and sensor networks to multilayered demand variables in a CG. Furthermore, we extend the CG-based framework by mapping different congestion mitigation strategies to CG layers individually or in combination, allowing the marginal effects and potential migration magnitudes of the strategies to be reliably quantified. Using the TensorFlow architecture, we evaluate our framework on the Sioux Falls network and present a large-scale case study based on a subnetwork of Beijing using a data set from the metropolitan planning organization.",
keywords = "computational graph, congestion mitigation, marginal analyses, TensorFlow, traffic demand estimation",
author = "Jianping Sun and Jifu Guo and Xin Wu and Qian Zhu and Danting Wu and Kai Xian and Xuesong Zhou",
year = "2019",
month = "5",
day = "15",
doi = "10.3390/s19102254",
language = "English (US)",
volume = "19",
journal = "Sensors (Switzerland)",
issn = "1424-8220",
publisher = "Multidisciplinary Digital Publishing Institute (MDPI)",
number = "10",

}

TY - JOUR

T1 - Analyzing the Impact of Traffic Congestion Mitigation

T2 - From an Explainable Neural Network Learning Framework to Marginal Effect Analyses

AU - Sun, Jianping

AU - Guo, Jifu

AU - Wu, Xin

AU - Zhu, Qian

AU - Wu, Danting

AU - Xian, Kai

AU - Zhou, Xuesong

PY - 2019/5/15

Y1 - 2019/5/15

N2 - Computational graphs (CGs) have been widely utilized in numerical analysis and deep learning to represent directed forward networks of data flows between operations. This paper aims to develop an explainable learning framework that can fully integrate three major steps of decision support: Synthesis of diverse traffic data, multilayered traffic demand estimation, and marginal effect analyses for transport policies. Following the big data-driven transportation computational graph (BTCG) framework, which is an emerging framework for explainable neural networks, we map different external traffic measurements collected from household survey data, mobile phone data, floating car data, and sensor networks to multilayered demand variables in a CG. Furthermore, we extend the CG-based framework by mapping different congestion mitigation strategies to CG layers individually or in combination, allowing the marginal effects and potential migration magnitudes of the strategies to be reliably quantified. Using the TensorFlow architecture, we evaluate our framework on the Sioux Falls network and present a large-scale case study based on a subnetwork of Beijing using a data set from the metropolitan planning organization.

AB - Computational graphs (CGs) have been widely utilized in numerical analysis and deep learning to represent directed forward networks of data flows between operations. This paper aims to develop an explainable learning framework that can fully integrate three major steps of decision support: Synthesis of diverse traffic data, multilayered traffic demand estimation, and marginal effect analyses for transport policies. Following the big data-driven transportation computational graph (BTCG) framework, which is an emerging framework for explainable neural networks, we map different external traffic measurements collected from household survey data, mobile phone data, floating car data, and sensor networks to multilayered demand variables in a CG. Furthermore, we extend the CG-based framework by mapping different congestion mitigation strategies to CG layers individually or in combination, allowing the marginal effects and potential migration magnitudes of the strategies to be reliably quantified. Using the TensorFlow architecture, we evaluate our framework on the Sioux Falls network and present a large-scale case study based on a subnetwork of Beijing using a data set from the metropolitan planning organization.

KW - computational graph

KW - congestion mitigation

KW - marginal analyses

KW - TensorFlow

KW - traffic demand estimation

UR - http://www.scopus.com/inward/record.url?scp=85066267536&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85066267536&partnerID=8YFLogxK

U2 - 10.3390/s19102254

DO - 10.3390/s19102254

M3 - Article

C2 - 31096706

AN - SCOPUS:85066267536

VL - 19

JO - Sensors (Switzerland)

JF - Sensors (Switzerland)

SN - 1424-8220

IS - 10

ER -