A route planner must try to schedule the deliveries by a fleet of vehicles such that customer requirements are met and management objectives are satisfied. In most cases, the number of feasible arrangements is legion, and calculations relating to individual vehicle loads, mileages, delivery times, etc, are tedious, allowing only a small fraction of possible route plans to be established and compared. The problem presents an ideal opportunity for computer application, not least to ensure that solutions are timely and error free. Several algorithms have been developed to improve the quality of vehicle routes, but in practice only those that rely on simple selection rules have found widespread acceptance, due to the innate complexity of the calculations that follow from a more rigorous approach and to the great variety of customer, vehicle, and operational characteristics that distinguish transport systems and which must be accommodated. The method presented here is based upon the well known "savings" criterion, but avoids many of its deficiencies by employing a random selection mode and producing (efficiently) a large sample of schedules from which to choose the most suitable. In particular, this allows greater flexibility in defining management objectives, and has led to substantial reductions in both fleet sizes and distances travelled, compared to published results, for a set of nine test cases each involving more than 200 customer locations. (Author/TRRL)

  • Availability:
  • Corporate Authors:

    Deakin University

    School of Management
    Geelong, Victoria,   Australia 
  • Authors:
    • Buxey, G
  • Publication Date: 1985-7

Media Info

  • Features: References;
  • Pagination: 16 p.
  • Serial:

Subject/Index Terms

Filing Info

  • Accession Number: 00491487
  • Record Type: Publication
  • Source Agency: ARRB
  • ISBN: 0-86828-399-1
  • Report/Paper Numbers: No. 78
  • Files: ITRD, TRIS, ATRI
  • Created Date: Mar 31 1990 12:00AM