The model of Ben-Akiva, Cyna, and de Palma is extended to represent trip departure time and route choice decisions when total demandis elastic. The simple case treated has two parallel routes with travelers jointly selecting route and departure time. The delays are assumed to occur at bottlenecks of limited capacity (bridge, tunnel, etc) and a simple queueing model is employed to determine waiting time as a function of queue length at the time of arrival at the end of the queue. The day-to-day adjustment of the distribution of traffic is derived from a dynamic Markovian model. Numerical simulation experiments are performed to demonstrate the possible changes in the pattern of peak-period congestion when capacity of a bottleneck is changed. The results demonstrate some of the interdepencies that may exist among different bottlenecks in a road network. It is shown, in particular, that, in the presence of elastic demand, congestion may persist even when capacity of a bottleneck is expanded to meet the highest level of existing traffic flows. This does not mean, however, that expanding the capacity of a bottleneck and thus diverting trips from other routes cannot be a successful strategy for reducing schedule delays and traffic congestion along other routes, if that is the objective of traffic management. In addition, it is shown that ifthe capacities of the bottlenecks remain constant on average, but fluctuate from day to day because of stochastic factors (such as weather conditions), average traffic delays tend to increase.
Abstract