Bayesian ranking and selection model for the Second-best Network Pricing Problem

Zhen Tan, H. Oliver Gao

Research output: Chapter in Book/Conference proceedingConference contributionpeer-review

2 Citations (Scopus)


We adopt a Bayesian ranking and selection (R&S) model to solve the Second-best Network Pricing Problem (SNPP) in transportation. The objective of SNPP is to find an optimal subset of links and toll levels so as to minimize the total travel time on the network. It is an NP-hard problem with a large number of candidate solutions. We consider every combination of tollable link(s) and toll levels as an 'alternative', and the problem's objective function value is regarded as a 'reward', with uncertainties modeled by normal perturbations to the travel demand. We use a linear belief based Knowledge Gradient sampling policy to maximize the expected reward, with Monte Carlo sampling of the hyperparameters used to reduce the choice set size. Simulation experiments for a benchmark network show the effectiveness of the proposed method and its superior performance to a Sample Average Approximation based Genetic Algorithm.

Original languageEnglish
Title of host publication2016 Winter Simulation Conference
Subtitle of host publicationSimulating Complex Service Systems, WSC 2016
EditorsTheresa M. Roeder, Peter I. Frazier, Robert Szechtman, Enlu Zhou
PublisherInstitute of Electrical and Electronics Engineers Inc.
Number of pages12
ISBN (Electronic)9781509044863
Publication statusPublished - 2 Jul 2016
Externally publishedYes
Event2016 Winter Simulation Conference, WSC 2016 - Arlington, United States
Duration: 11 Dec 201614 Dec 2016

Publication series

NameProceedings - Winter Simulation Conference
ISSN (Print)0891-7736


Conference2016 Winter Simulation Conference, WSC 2016
Country/TerritoryUnited States

ASJC Scopus subject areas

  • Software
  • Modelling and Simulation
  • Computer Science Applications


Dive into the research topics of 'Bayesian ranking and selection model for the Second-best Network Pricing Problem'. Together they form a unique fingerprint.

Cite this