A lossless spatial aggregation procedure for a class of capacity constrained traffic assignment models incorporating point queues

Abstract

In this paper two novel spatial aggregation procedures are proposed. A network aggregation procedure based on a travel time delay decomposition method and a zonal aggregation procedure based on a path redistribution scheme. The effectiveness of these procedures lies in the fact that they, unlike existing aggregation methods, exploit available information regarding the application context and the characteristics of the adopted traffic assignment procedure. The context considered involves all applications that require path and inter-zonal travel times as output. A typical example of such applications are quick-scan methods, which have become increasing popular in recent years. The proposed procedures are compatible with a class of traffic assignment procedures incorporating (residual) point queues. Furthermore, one can choose to combine network aggregation with zonal aggregation to increase the effectiveness of the procedure. Results are demonstrated via theoretical examples as well as a large-scale case study. In the case study it is shown that network loading times can be reduced to as little as 4% of the original situation without suffering any information loss.