DISTANCE-BASED MODEL FOR ESTIMATING A BUS ROUTE ORIGIN-DESTINATION MATRIX

Abstract

An origin-destination (O-D) is a valuable tool for bus service planning. Unfortunately this trip table is not commonly available to the planner because of survey costs. However, stop-level on-off totals are often available. Past research has concentrated on using these totals along with a small O-D survey as a matrix to generate the full O-D matrix. Such a seed is subject to bias and sampling error and also incurs the survey cost. A method is described in which the seed is generated using a propensity function that models the propensity of travel as a function of travel distance. The proposed function is a product of a power term and an exponential term, equivalent to a gamma distribution. When applied to one-directional travel, the gamma seed is shown to be reduced to a power function. The power function exponent is estimated by maximum likelihood for data from bus routes in Boston and Miami and is consistently found to be near 1.0. The gamma seed combined with the biproportional method to match origin and destination totals is shown to be effective in generating O-D matrixes for Boston and Miami routes. In a practical design application, design measures were found to be relatively insensitive to changes in the function parameter.

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 — 12 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.