Beyond Political Boundaries: Constructing Network Models for Megaregion Planning

The scale of urban planning is now focusing on megaregions in addition to metropolitan areas and states. The traffic assignment problem (TAP), used to study traffic flow patterns on networks, is a crucial step in urban planning. Megaregional networks transcend planning agency jurisdictions, challenging current network models and computational resources. This study aims to solve TAP on a megaregional scale by applying an algorithm based on the decomposition approach for the static TAP (DSTAP) that uses network decompositions based on network geography. In the first part of this research, the authors compare two partitioning algorithms for finding network partitions for megaregions by minimizing the number of subnetwork boundary nodes and the time required to solve DSTAP. The flow-based spectral partitioning generates flow balanced subnetworks which reduce the per iteration computation time and lead to faster convergence compared to the agglomerative partitioning algorithm. In the second part of this research, the authors propose a decomposition heuristic for large scale networks, allowing parallelization of TAP. The heuristic reduces the computational time for DSTAP by simplifying interactions within the subnetwork. For the uncongested Texas network, the proposed heuristic led to marginal 5% savings in computational time than state-of-the-art TAP methods, while for the congested scenario, the heuristic observed about 70% savings in computation time for the same convergence level. However, the heuristic leads to a lower bound on the relative gap value at termination (called heuristic gap) which ranges between 9E-3 and 5E-4 for the experiments conducted on the Texas statewide network.


  • English

Media Info

  • Media Type: Digital/other
  • Edition: Technical Report
  • Features: Figures; Maps; References; Tables;
  • Pagination: 50p

Subject/Index Terms

Filing Info

  • Accession Number: 01691027
  • Record Type: Publication
  • Report/Paper Numbers: CM2-11
  • Contract Numbers: 69A3551747135
  • Created Date: Jan 10 2019 3:26PM