PARTIAL MATRICES, EMPIRICAL DETERRENCE FUNCTIONS AND ILL-DEFINED RESULTS

This paper is concerned with the problem of fitting the "trip-distribution" stage of a transportation model to a set of observed data, composed of matrices of trips and generalised costs. Reference is made to a method proposed by other research workers involving the use of partial matrix techniques, and this term is defined as implying the use of empirical deterrence functions of generalised costs. The method used by the authors is discussed by means of an example related to a 9-zone system. Calibration results are presented in a series of tables, the results analysed and discussed, and consideration given to theory, function components, diagnosis and treatment. The authors conclude that it is possible to reproduce exactly a set of 'observed' trips, even when the calibrated factors are not unique to the extent usually assumed in empirical function fitting. This lack of uniqueness is considered irrelevant to the reproduction of the original data, but that when the condition is present, the model can be used neither for filling out a partial matrix nor for predicting the effects of cost-changes on trips. (TRRL)

  • Availability:
  • Corporate Authors:

    Printerhall Limited

    29 Newmart Street
    London W1P 3PE,   England 
  • Authors:
    • Day, MJL
    • Hawkins, A F
  • Publication Date: 1979-8-9

Media Info

Subject/Index Terms

Filing Info

  • Accession Number: 00310282
  • Record Type: Publication
  • Source Agency: Transport Research Laboratory
  • Files: ITRD, TRIS
  • Created Date: Oct 27 1980 12:00AM