A Cell-Based Many-to-One Dynamic System Optimal Model and Its Heuristic Solution Method for Emergency Evacuation

An efficient prescriptive dynamic traffic assignment model is critical for the effective traffic management under emergency evacuation. Although a number of dynamic traffic assignment models have been proposed in the previous studies, it is almost impossible to apply them for real-time emergency traffic management due to the high computational cost. However, for the real-time emergency traffic management, computational efficiency becomes essential but reasonably detailed representation of traffic flow dynamics has to be maintained. By assuming that all evacuees have only one destination, the safe area, the prescriptive model is established as a many-to-one dynamic system optimal problem and a linear program is formulated based on the cellular representation of traffic network. To achieve the required computational efficiency, instead of solving the linear program directly, we develop a heuristic algorithm called HASTE and provide a close approximation to the optimal solution. We apply the system optimal model to a hypothetical emergency evacuation scenario in the downtown of Minneapolis, Minnesota, and the solutions are generated both by the linear program solver and by HASTE. The comparisons between different solution approaches indicate that the heuristic can provide close-to-optimal evacuation strategies with a much higher computational efficiency, which is important to real-time evacuation operations.

Language

  • English

Media Info

  • Media Type: CD-ROM
  • Features: Appendices; Figures; Maps; References; Tables;
  • Pagination: 20p
  • Monograph Title: TRB 86th Annual Meeting Compendium of Papers CD-ROM

Subject/Index Terms

Filing Info

  • Accession Number: 01047485
  • Record Type: Publication
  • Report/Paper Numbers: 07-2261
  • Files: TRIS, TRB
  • Created Date: Feb 8 2007 6:56PM