Synthesis of Traveler Choice Research: Improving Modeling Accuracy for Better Transportation Decisionmaking

Model Integration

As previously discussed, the development of advanced travel behavior and demand models and the development of transport supply models have been relatively independent of one another. Thus, the demand and supply models are each formulated to use forecast outputs from the other model without feedback. As a result, the level of service input to demand models is not necessarily the same as that output from the supply models nor is the demand input to supply models necessarily the same as that output from the demand models. Thus, it is important to integrate demand and supply models to ensure consistency between supply and demand. Efforts to integrate demand and supply models have been made in recent years. Lam and Huang presented a mathematical formulation for dynamic traffic assignment for modeling simultaneous location, route, and departure time choices.⁽¹¹⁴⁾ Their model can be used as a simplified travel demand analysis tool but cannot capture travel behavior complexity. Lin et al. proposed an integration of an activity-based model simulator with a dynamic traffic assignment model, where feedback convergence is measured by aggregated travel time and number of trips.⁽¹¹⁵⁾ The authors showed that the initial differences are substantial and can be reduced dramatically on an aggregated level. Lu and Mahmassani presented a “joint route and departure time network equilibrium assignment model explicitly considering heterogeneous users with different preferred arrival times at destinations, values of time, and values of early and late schedule delays.”⁽¹¹⁶⁾ Their “multicriterion simultaneous route and departure time user equilibrium” determined both changes in route choice in response to dynamic pricing as well as temporal shifts toward less congested periods. Application of this model in a large network setting remains limited by computational capabilities, but the model can realistically be applied to alleviate congestion by finding optimal dynamic pricing schemes by location, pricing periods, and toll charges.

On a completely disaggregate level, some agent-based simulations also incorporate demand and supply models. Esser and Nagel and Rieser et al. developed a multiagent microsimulation module that integrates activity generation, route assignment, and network loading.^(117,118) Generated daily activity-based plans are executed by agents and assigned to the route. The traffic simulation is then used to evaluate those activity plans. Changes in activity start times, route choices, etc. are randomly adjusted and high-scoring activity plans are executed.

Although the need for integration of transportation demand and supply has been accepted for years, much of the research is still in the conceptual stage.