Statistical metamodeling of dynamic network loading

Wenjing Song, Ke Han, Yiou Wang, Terry Friesz, Enrique Del Castillo

Research output: Contribution to journalArticle

7 Citations (Scopus)

Abstract

Dynamic traffic assignment models rely on a network performance module known as dynamic network loading(DNL), which expresses the dynamics of flow propagation, flow conservation, and travel delay at a network level. The DNL defines the so-called network delay operator, which maps a set of path departure rates to a set of path travel times. It is widely known that the delay operator is not available in closed form, and has undesirable properties that severely complicate DTA analysis and computation, such as discontinuity, non-differentiability, non-monotonicity, and computational inefficiency. This paper proposes a fresh take on this important and difficult problem, by providing a class of surrogate DNL models based on a statistical learning method known as Kriging. We present a metamodeling framework that systematically approximates DNL models and is flexible in the sense of allowing the modeler to make trade-offs among model granularity, complexity, and accuracy. It is shown that such surrogate DNL models yield highly accurate approximations (with errors below 8%) and superior computational efficiency (9 to 455 times faster than conventional DNL procedures). Moreover, these approximate DNL models admit closed-form and analytical delay operators, which are Lipschitz continuous and infinitely differentiable, while possessing closed-form Jacobians. The implications of these desirable properties for DTA research and model applications are discussed in depth.

Original languageEnglish (US)
Pages (from-to)263-282
Number of pages20
JournalTransportation Research Procedia
Volume23
DOIs
StatePublished - Jan 1 2017

Fingerprint

Differential thermal analysis
travel
Travel time
Network performance
Computational efficiency
Telecommunication traffic
learning method
Conservation
conservation
traffic
efficiency
performance
time

All Science Journal Classification (ASJC) codes

  • Transportation

Cite this

@article{94cf513ae535449295e1a9c1b22d731e,
title = "Statistical metamodeling of dynamic network loading",
abstract = "Dynamic traffic assignment models rely on a network performance module known as dynamic network loading(DNL), which expresses the dynamics of flow propagation, flow conservation, and travel delay at a network level. The DNL defines the so-called network delay operator, which maps a set of path departure rates to a set of path travel times. It is widely known that the delay operator is not available in closed form, and has undesirable properties that severely complicate DTA analysis and computation, such as discontinuity, non-differentiability, non-monotonicity, and computational inefficiency. This paper proposes a fresh take on this important and difficult problem, by providing a class of surrogate DNL models based on a statistical learning method known as Kriging. We present a metamodeling framework that systematically approximates DNL models and is flexible in the sense of allowing the modeler to make trade-offs among model granularity, complexity, and accuracy. It is shown that such surrogate DNL models yield highly accurate approximations (with errors below 8{\%}) and superior computational efficiency (9 to 455 times faster than conventional DNL procedures). Moreover, these approximate DNL models admit closed-form and analytical delay operators, which are Lipschitz continuous and infinitely differentiable, while possessing closed-form Jacobians. The implications of these desirable properties for DTA research and model applications are discussed in depth.",
author = "Wenjing Song and Ke Han and Yiou Wang and Terry Friesz and {Del Castillo}, Enrique",
year = "2017",
month = "1",
day = "1",
doi = "10.1016/j.trpro.2017.05.016",
language = "English (US)",
volume = "23",
pages = "263--282",
journal = "Transportation Research Procedia",
issn = "2352-1457",
publisher = "Elsevier BV",

}

Statistical metamodeling of dynamic network loading. / Song, Wenjing; Han, Ke; Wang, Yiou; Friesz, Terry; Del Castillo, Enrique.

In: Transportation Research Procedia, Vol. 23, 01.01.2017, p. 263-282.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Statistical metamodeling of dynamic network loading

AU - Song, Wenjing

AU - Han, Ke

AU - Wang, Yiou

AU - Friesz, Terry

AU - Del Castillo, Enrique

PY - 2017/1/1

Y1 - 2017/1/1

N2 - Dynamic traffic assignment models rely on a network performance module known as dynamic network loading(DNL), which expresses the dynamics of flow propagation, flow conservation, and travel delay at a network level. The DNL defines the so-called network delay operator, which maps a set of path departure rates to a set of path travel times. It is widely known that the delay operator is not available in closed form, and has undesirable properties that severely complicate DTA analysis and computation, such as discontinuity, non-differentiability, non-monotonicity, and computational inefficiency. This paper proposes a fresh take on this important and difficult problem, by providing a class of surrogate DNL models based on a statistical learning method known as Kriging. We present a metamodeling framework that systematically approximates DNL models and is flexible in the sense of allowing the modeler to make trade-offs among model granularity, complexity, and accuracy. It is shown that such surrogate DNL models yield highly accurate approximations (with errors below 8%) and superior computational efficiency (9 to 455 times faster than conventional DNL procedures). Moreover, these approximate DNL models admit closed-form and analytical delay operators, which are Lipschitz continuous and infinitely differentiable, while possessing closed-form Jacobians. The implications of these desirable properties for DTA research and model applications are discussed in depth.

AB - Dynamic traffic assignment models rely on a network performance module known as dynamic network loading(DNL), which expresses the dynamics of flow propagation, flow conservation, and travel delay at a network level. The DNL defines the so-called network delay operator, which maps a set of path departure rates to a set of path travel times. It is widely known that the delay operator is not available in closed form, and has undesirable properties that severely complicate DTA analysis and computation, such as discontinuity, non-differentiability, non-monotonicity, and computational inefficiency. This paper proposes a fresh take on this important and difficult problem, by providing a class of surrogate DNL models based on a statistical learning method known as Kriging. We present a metamodeling framework that systematically approximates DNL models and is flexible in the sense of allowing the modeler to make trade-offs among model granularity, complexity, and accuracy. It is shown that such surrogate DNL models yield highly accurate approximations (with errors below 8%) and superior computational efficiency (9 to 455 times faster than conventional DNL procedures). Moreover, these approximate DNL models admit closed-form and analytical delay operators, which are Lipschitz continuous and infinitely differentiable, while possessing closed-form Jacobians. The implications of these desirable properties for DTA research and model applications are discussed in depth.

UR - http://www.scopus.com/inward/record.url?scp=85020516663&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85020516663&partnerID=8YFLogxK

U2 - 10.1016/j.trpro.2017.05.016

DO - 10.1016/j.trpro.2017.05.016

M3 - Article

AN - SCOPUS:85020516663

VL - 23

SP - 263

EP - 282

JO - Transportation Research Procedia

JF - Transportation Research Procedia

SN - 2352-1457

ER -