Long-term scheduling for road network disaster recovery

In this paper, the authors address the problem of scheduling reconstruction efforts in a transport network struck by a disaster. The recovery of the road network is assumed to require a substantial amount of resources, e.g. budget, working crew, equipment; which are limited in availability. The authors assume that the recovery of the road network will take place over a pre-defined planning horizon divided into time periods. Further, the authors assume that the time periods are long enough so that travelers have the time to adjust their route choice preferences. The authors formulate this network recovery problem as a bilevel optimization problem wherein travelers are assumed to behave under user equilibrium conditions. The authors propose an exact enumerative approach that builds on prior work and requires the solution of an exponential number of traffic equilibrium problems, as well as three heuristics that are employed from the scheduling literature. Numerical experiments are designed on a realistic transport network wherein two disaster scenarios are considered. Both disaster scenarios aim to represent the possible damage caused by disasters that exhibit a fault-line topology, such as earthquakes. The numerical results obtained highlight the performance of the proposed heuristics for the bilevel network recovery scheduling problem at hand, but also reveal that these methods may fall short of optimal solutions when the amount of recovery resources is low compared to the demand of recovery projects.

Language

  • English

Media Info

Subject/Index Terms

Filing Info

  • Accession Number: 01722262
  • Record Type: Publication
  • Files: TRIS
  • Created Date: Nov 14 2019 9:30AM