Empirical studies of work trip distribution models: Fid-Bau Portal

Empirical studies of work trip distribution models

Sikdar, P.K. / Hutchinson, B.G.

Abstract The goodness of fit characteristics of three members of the Wilson-family of gravity models in explaining observed work trip interchanges are examined using an almost 100% sample of journey to work trips in Edmonton, Alberta. Model parameters of the three model types and several variants are estimated for four zone systems ranging in numbers from 22 to 234 zones. Changes in parameter magnitudes and model performance are examined across model types and spatial scales. It is concluded that current gravity models calibrated to cross-sectional data and using only trip end and aggregate travel cost constraints have inadequate explanatory power to justify their use in planning studies. Additional trip end constraints, such as socio-economic stratification, do not yield significant improvements and trip interchange constraints appear to offer the most productive improvement possibilities. The most important factor which is not captured by current cross-sectional models is the timing of urban development which can only reasonably be captured through trip interchange constraints.