Discrete choice, random utility models, such as the multinomial logit and multinomial probit models, provide a convenient behavioral and mathematical framework for carrying out travel demand analysis. When these models first were brought to the attention of the transportation community in the early 1970s, it was argued they had several important theoretical and practical advantages over other available travel demand modelling approaches such as the familiar four-step process. The advantages claimed for random utility models included their basis in an explicit principle of human behavior, their ability to treat an unusually wide range of travel choices and policy variables, and their ability to make efficient use of data. It was argued that these advantages enable random utility models to forecast travel more accurately and less expensively than do other models. Since the early 1970s, there has been much research and experience with random utility travel demand models, and these models are understood much better now than they were 10 years ago. The purpose of this paper is to review current knowledge of random utility modeling that affects practical travel demand analysis and to identify the implications of this knowledge for parctice. For the covering abstract of the conference see IRRD 276520. (Author/TRRL)

  • Supplemental Notes:
    • Transportation Planning Research Colloquium 1983 held in Zandvoort on December 14-16, 1983. Volume II.
  • Corporate Authors:

    Colloguium Vervoersplanologisch Speurwerk

    P.O. Box 45
    Delft,   Netherlands 
  • Authors:
    • Horowitz, J L
  • Publication Date: 1984

Media Info

Subject/Index Terms

Filing Info

  • Accession Number: 00389298
  • Record Type: Publication
  • Source Agency: Institute for Road Safety Research, SWOV
  • Files: ITRD, TRIS
  • Created Date: Oct 30 1984 12:00AM