Transportation network models and algorithms considering directional delay and prohibition for intersection movement

Applying the link-label technique enables to simultaneously take into consideration two links (or three nodes) in the path search process of the shortest path algorithms. The advantages of the link-label are: First the network expansion is required for representation of directional movements, thus network-building complexity is reduced and modification of network is flexible. Second the optimum solution is guaranteed since the optimality condition of the dynamic programming is satisfied. Third the vine-typed travel patterns such as U-Turn and P-turn frequently witnessed in urban road network can be reasonably explained as well as implemented.; This study aims to build theoretical background to exploit the advantages of the link-label in the transportation network analysis related models and algorithms, which have been developed based on the node-label technique. The main idea of this study is to develop methodologies to take into consideration directional delays and prohibitions lay between two adjacent links centering intersections. The concept is supported by the principle of reasonable path choice saying, "When a path is represented by a sequence of node, any node can be appeared more than once, while by link, any link should be appeared no more than once". User-optimal traffic assignments and multiple paths finding algorithms are research objects.; For applying the link-label technique to the user-optimal traffic assignment, first the Wordrop's route choice condition is redefined in terms of visited directional flows and all directional flows with embedded directional delay term. Second using the derived route choice conditions, the formulations and solution algorithms are proposed based on nonlinear programming (NLP), variational inequality (VI), and variational equality (VE) theories.; For applying the link-label technique to the multiple paths finding solution algorithms, three multiple paths finding algorithms, which were developed by Yen and Shier, are modified: the one-to-one algorithm by Yen and two one-to-all algorithms by Shier; the label setting and the label correcting algorithms.; Various case studies are demonstrated for validation and improvement of the proposed models and algorithms.