OPTIMAL DISPATCHING POLICIES BY DYNAMIC PROGRAMMING

The paper describes methods of determining optimal vehicle dispatching schedules by the use of dynamic programming techniques. Using cost criteria based on minimizing a combination of passenger delay and system capacity, these techniques were applied to vehicle scheduling for three variations of linear networks of dispatching stations: point-to-point (one way), line of stations (one way), and two-station line (round trip). FORTRAN programs were written to aid in both the generation and analysis of the optimal schedules. Various dispatching policies are examined with respect to system parameters such as vehicle capacity, load factor, and fleet size. An analysis and comparison of the optimal schedules in terms of passenger delay and vehicle fleet size are made with some non-optimal schedules similar to those used in many present-day operations. Optimal schedules yielding minimum passenger delay are shown to be superior with respect to most other system variables. The value of dynamic programming in these and future scheduling studies is discussed. (Author)

  • Corporate Authors:

    Massachusetts Institute of Technology

    Department of Civil Engineering, 77 Massachusetts Avenue
    Cambridge, MA  USA  02139
  • Authors:
    • Ward, D E
  • Publication Date: 1966-11-1

Media Info

  • Pagination: 102 p.

Subject/Index Terms

Filing Info

  • Accession Number: 00039089
  • Record Type: Publication
  • Source Agency: National Technical Information Service
  • Report/Paper Numbers: R66-55 Res Rpt
  • Contract Numbers: C-85-65t
  • Files: TRIS
  • Created Date: Nov 24 1973 12:00AM