Trip Table Realization: Underlying Stochasticity and Impacts on Assigned Link Flows

Uncertainty analyses in transportation systems and traffic assignment practices seem to have overlooked the stochasticity feature of a trip table, which in this paper refers to the day-to-day random variation in travel demand distribution. The authors consider the stochasticity is an inherent nature of an original trip table that is obtained from travel demand models or trip surveys, and try to gauge the imposed variations of link flows and costs. By using a Poisson trip table realization procedure, two different original trip tables from Chicago and Barceloa were “rehabilitated” to thirty stochastic realized trip tables. A series of comparisons were made between the original and realized trip tables, assigned link flows and costs, VMTs and VHTs. Absolute and relative variations were also measured among the thirty realized trip tables. The results suggested that although large discrepancy exists for the cell-level OD trips the overall variability of the assigned link flows and costs is fairly small. At the same time, large flow variations do exist on some links. The findings justified the common practice of only using an original trip table to do trip assignment, given the purpose of the analysis is to obtain overall network performance measurements, such as VMT, VHT. On the other hand, it should be cautioned in drawing conclusions on a subnetwork level analysis (individual link level) and scenario analysis where large link flow variations may be found. In such cases, the effect of trip table stochasticity should not be ignored and using multiple realized trip tables is recommended.