Network trip assignment and aggregation procedures in urban transportation planning and design

This study documents a performed research on transportation trip (traffic) assignment procedures as well as on techniques of aggregation applied to urban transportation networks. It should be seen as a contribution to the analysis and research on assignment and aggregation principles. It concludes with a procedure proposed in the hope of reaching a working and feasible transportation network level of aggregation.;Two aggregate networks consisting of 47 zones (295 links) and 183 zones (1,292 links) were used to investigate trip assignment procedures using the BPR link performance function. This function uses the free flow travel time and the volume/capacity ratio raised to the fourth power in order to compute for congestion. The major concern is, first, to study the applicability of three transportation network assignment algorithms: (1) All-or-nothing incremental, tree-by-tree, and Volume/capacity restrained. (2) Stochastic multi-path incremental, using double-pass method. (3) User-equilibrium using the convex combination method.;These algorithms simulate and assign traffic along (i) primary roads, and (ii) primary and secondary roads. Some effort was made to explain the underlying theories of the applied procedures and their equation structure.;The second major concern was to study the applicability of two transportation network aggregation procedures: (1) Abstraction aggregation as presented by Chan 1976. (2) Extraction aggregation as presented by Haghani and Daskin 1983. The User Equilibrium is used to simulate and to assign traffic on the aggregate networks where it proved to better assign flows on links. Some underlying theories of aggregation in different fields were explained and a review of related literature provided. The effects of the level of aggregation on assignment outcomes were investigated.;It was shown that Chan's abstraction aggregation could not be applied to certain types of transportation network, while Haghani and Daskin's extraction aggregation is not practical for large problems and has some deficiencies.;The City of Eindhoven in the Province of Brabant in The Netherlands was used for this study. Actual street counts, trip tables, and a survey on travel times and trips were characteristically available. The used network consists primarily of 1,286 zones and 6,648 links with 29,287 inter-zonal and 143 intra-zonal hourly trips.