An evolutionary game model and MCMC estimation for analyzing stochastic properties of traffic state on a road network

Yuki Shittaka, Takeshi Nagae

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)

Abstract

This paper proposes a novel methodology for evaluating the travel time reliability of an urban road network using latest advances in the evolutionary game theory and MCMC (Markov chain Monte Carlo). Because of the stochastic variation of congestion, the importance of travel time reliability on an urban road network is attractive for research. Moreover, analyzing the stochastic properties of travel time is essential for maintaining an urban road network reliable. Our proposed method has the following notable features: (i) it obtains a stationary distribution of traffic flow patterns which stems from a stochastic day-to-day dynamics of a population with rational users; (ii) this distribution is consistent with the traditional SUE (stochastic user equilibrium), and it is adaptable to an economical evaluation framework including cost-benefit analysis of road development projects that is standardized by government; and (iii) it can estimate the 95 percentile of minimum travel time of each OD pair for a practical size network. We first use an evolutionary game model proposed by Sandholm (2009) to describe the stochastic day-to-day dynamics of the traffic flow patterns generated by the perturbed best response of each user. This model is characterized by a stationary distribution of traffic flow patterns, and it is consistent with the SUE. We then develop a sampling method to generate sample traffic flow patterns from the stationary distribution by applying the MCMC re-sampling method proposed by Wei at el. (2012). Finally, we develop a numerical method for estimating the 95 percentile of minimum travel time efficiently applying the importance sampling and a randomized algorithm. Several numerical applications of the proposed method are demonstrated on a moderate size network.

Original languageEnglish
Title of host publicationProceedings - 2016 International Conference on Agents, ICA 2016
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages82-85
Number of pages4
ISBN (Electronic)9781509039319
DOIs
Publication statusPublished - 2017 Jan 10
Event1st IEEE International Conference on Agents, ICA 2016 - Matsue, Shimane, Japan
Duration: 2016 Sep 282016 Sep 30

Other

Other1st IEEE International Conference on Agents, ICA 2016
CountryJapan
CityMatsue, Shimane
Period16/9/2816/9/30

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Hardware and Architecture
  • Modelling and Simulation

Fingerprint Dive into the research topics of 'An evolutionary game model and MCMC estimation for analyzing stochastic properties of traffic state on a road network'. Together they form a unique fingerprint.

  • Cite this

    Shittaka, Y., & Nagae, T. (2017). An evolutionary game model and MCMC estimation for analyzing stochastic properties of traffic state on a road network. In Proceedings - 2016 International Conference on Agents, ICA 2016 (pp. 82-85). [7812972] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICA.2016.22