Modeling the morning commute for urban networks with cruising-for-parking: An MFD approach

Wei Liu, Nikolas Geroliminis

Research output: Journal article publicationJournal articleAcademic researchpeer-review

95 Citations (Scopus)


This study focuses on the morning commute problem with explicit consideration of cruising-for-parking, and its adverse impacts on traffic congestion. The cruising-for-parking is modeled through a dynamic aggregated traffic model for networks: the Macroscopic Fundamental Diagram (MFD). Firstly, we formulate the commuting equilibrium in a congested downtown network where travelers have to cruise for curbside parking spaces. The cruising-for-parking would yield longer trip distance and smaller network outflow, and thus can induce severe congestion and lengthen the morning peak. We then develop a dynamic model of pricing for the network to reduce total social cost, which includes cruising time cost, moving time cost (moving or in-transit time, which is the duration during which vehicles move close to the destination but do not cruise for parking yet), and schedule delay cost. We show that under specific assumptions, at the system optimum, the downtown network should be operating at the maximum production of its MFD. However, the cruising effect is not fully eliminated. We also show that the time-dependent toll to support the system optimum has a different shape than the classical fine toll in Vickrey's bottleneck model. In the end, analytical results are illustrated and verified with numerical experiments.

Original languageEnglish
Pages (from-to)470-494
Number of pages25
JournalTransportation Research Part B: Methodological
Publication statusPublished - 1 Nov 2016
Externally publishedYes


  • Cruising-for-parking
  • MFD
  • Morning commute
  • Pricing

ASJC Scopus subject areas

  • Civil and Structural Engineering
  • Transportation


Dive into the research topics of 'Modeling the morning commute for urban networks with cruising-for-parking: An MFD approach'. Together they form a unique fingerprint.

Cite this