Started in the late 1970s, research into DTA, by representing time variations in traffic
flows and conditions, has tried to reflect the reality that traffic networks are generally not in a
steady state. To retain the advantages of an equilibrium approach, the notion of user equilibrium
needed to be extended in two ways. The first extension generalizes the static model’s perfect
traveler information assumption and route choice criterion, recognizing that travel times on
network links vary over time. Travelers are assumed to know or anticipate future travel
conditions along the journey (through learning from the past trials) and, in choosing an O-D
route, they are assumed to minimize the O-D travel time that they will actually experience; this
will depend on when they arrive at the various links along a route and on the travel times that
prevail on the links at those specific future times. (This is in contrast to a route evaluation
approach that considers only the link travel-times that prevail at the instant of departure from the origin. More discussion about experienced versus instantaneous travel times is provided in the
section Instantaneous and Experienced Travel Times.) Since travelers who depart from an origin
to a destination at different times will experience different travel times, the second extension
recognizes that, in a dynamic approach, the user equilibrium condition of equal travel times on
used routes applies only to travelers who are assumed to depart at the same time between the
same O-D pair. (An important generalization of DTA simultaneously determines travellers’
choice of departure time and route. This model can directly analyze phenomena such as peak
spreading in response to congestion dynamics or time-varying tolls. It is still the subject of active
research, and will not be considered further here.)