A Comparison of System Optimal and User Optimal Route Guidance.

The work described in this paper (carried out under the EC `DRIVE' programme) extends the simulations described in Working Paper 315, with the aim of studying the likely benefits to and reactions of drivers to system optimal (SO) route guidance - in particular, these effects are compared with those obtained under user optimal (UE) guidance. The model used is again one of a multiple user class equilibrium assignment, so that equipped drivers may be directed to more than one route per origin-destination movement. UE and SO guidance are compared, at different levels of equipped vehicles and demand levels, on the basis of the number of routes they recommend and the similarity of the flows on these routes, as well as link-based properties such as actual flows and queues resulting. These serve to demonstrate the extent to which the routes recommended under UE guidance serve as proxies to those under SO guidance. Secondly, a comparison is made of average (dis)benefits to guided drivers as well as the excess travel time incurred by individual equipped drivers in following SO, as opposed to UE guidance, in order to determine the extent of user sub-optimality of SO routing. Thirdly, input from a parallel DRIVE project, investigating user reactions to guidance information, is used to infer the extent to which drivers are likely to accept the sub-optimality of SO guidance, and the factors which are likely to influence their acceptance. Finally, some preliminary analysis is performed on combined strategies, which aim to strike a balance between the system benefits of SO guidance and the user benefits of UE routing.