Scheduled Hyperpath: A Strategy for Reliable Routing and Scheduling of Deliveries in Time-Dependent Networks with Random Delays

This paper presents a new method for improving the on-time arrival reliability of deliveries. A dispatcher wishes to minimize the total expected travel time and determines a set of attractive links before the trip. Drivers choose between them en route on the basis of the traffic conditions that they observe on immediate successor links. Such a set of attractive paths with associated departure time and decision rule collectively offers a better expected travel time than any single path on its own and is referred to as a "scheduled hyperpath." The hyperpath concept was exploited by drawing of a parallel between the wait for randomly arriving public transport services and the wait for available capacity in traffic networks with random delays. Adaptation of those methods to the just-in-time delivery context follows from established assertions that consideration of multiple paths and adaptive path selection in networks with random travel times offers shorter travel times than selection of an a priori path with the least expected travel time. The HyperDOT algorithm presented identifies an optimal scheduled hyperpath sequentially in reverse topological and chronological order by use of the historical distributions of link travel times for discrete time intervals as inputs. Two areas of application are discussed: multiple simultaneous deliveries and repetitive shipments. Although the proposed strategy is advantageous in networks with multiple delays, it may lead to the discarding of routes that remain the shortest even when they are delayed because of the myopic character of information about network conditions.

Language

  • English

Media Info

Subject/Index Terms

Filing Info

  • Accession Number: 01477167
  • Record Type: Publication
  • ISBN: 9780309286978
  • Report/Paper Numbers: 13-3696
  • Files: TRIS, TRB, ATRI
  • Created Date: Apr 1 2013 10:50AM