This paper considers the modeling of road congestion subject to peak-load demand. The standard model contains ambiguities and is poorly specified. These problems can be eliminated by working with a structural model that explicitly treats the congestion technology and drivers' behavioral decisions. The paper provides a detailed analysis of a particular structural model--William Vickrey's model of bottleneck congestion in the morning rush-hour auto commute, extended to treat elastic (i.e., price-sensitive) demand--and examines some economic implications of the structural approach.

We consider an individual who lacks the information-processing capacity required for a direct comparison of all feasible allocations. Instead of finding at once a best allocation, the individual myopically adjusts his current allocation toward higher utility. The individual makes adjustment errors inversely proportional to his ability to choose. We compare the stationary state of this process with the standard model. We see how an imperfect ability to choose modifies both positive and normative predictions of the standard model and how the standard model can be obtained from our more general one as the special case corresponding to perfect ability.

This paper considers the modeling of road congestion subject to peak-load demand. The standard model contains ambiguities and is poorly specified. These problems can be eliminated by working with a structural model that explicitly treats the congestion technology and drivers' behavioral decisions. The paper provides a detailed analysis of a particular structural model--William Vickrey' s model of bottleneck congestion in the morning rush-hour auto commute extended to treat elastic (i.e., price-sensitive) demand--and examin es some economic implications of the structural approach.

The authors consider an individual who lacks the information-processing capacity required for a direct comparison of all feasible allocations. Instead of finding at once a best allocation, the individual myopically adjusts his current allocation toward higher utility. The individual makes adjustment errors inversely proportional to his ability to choose. The authors compare the stationary state of this process with the standard model. They see how an imperfect ability to choose modifies both positive and normative predictions of the standard model and how the standard model can be obtained from the authors' more general one as the special case corresponding to perfect ability.