ESTIMATION OF ORIGIN-DESTINATION MATRICES WITH CONSTRAINED REGRESSION

Abstract

The use of constrained generalized leastsquares (CGLS) regression to estimate origin-destination travel matrices from aggregate data is described. The CGLS method does not require general surveys but allows any available data to be included. Variances of matrix entry estimates can be estimated and used as measures of uncertainty or to suggest additional sampling strategy. Two case studies are described from applications to data from Portland, Oregon. The first involves expanding a matrix of transit work trips to all transit trips. Second, a gravity-type model of trip distribution for all work trips is estimated. Comparisons are made with other estimation methods with respect to accuracy, computational effort, and the use of uncertainty measures.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 24 citations exist upstream.