ESTIMATING PARAMETERS IN MODELS OF GAP ACCEPTANCE BY QUEUES

This paper compares five variants of the least squares method of the number of vehicles merging into a gap on a major road with a proposed alternative maximum likelihood method. Analysis shows that three of the five least squares methods are seriously biased. The other two, though not seriously biased, are relatively inefficient compared with the maximum likelihood method. The bias can be reduced by rejecting those gaps into which no vehicles entered, but much of the data then becomes redundant. Evidence is given to suggest that the relation between gap length and the number of vehicles entering it is nonlinear and the second vehicle in a queue has a critical gap approximately 0.75 to 1.0 seconds less than that for other vehicles. The authors claim that the least squares method is often an inappropriate method but is used because it is a conveniently available package on most computers. Maximum likelihood estimates are unbiased and use the most efficient available in terms of mean squared error; there is a significant gain in efficiency when compared with any of the least squares method. (TRRL)

  • Availability:
  • Corporate Authors:

    Printerhall Limited

    29 Newmart Street
    London W1P 3PE,   England 
  • Authors:
    • MAHER, M J
    • Dowse, R J
  • Publication Date: 1983-5

Media Info

Subject/Index Terms

Filing Info

  • Accession Number: 00381004
  • Record Type: Publication
  • Source Agency: Transport Research Laboratory
  • Report/Paper Numbers: HS-035 414
  • Files: HSL, ITRD, TRIS
  • Created Date: Jan 30 1984 12:00AM