A day-to-day and within-day dynamic stochastic assignment model

• Published in: Transportation research. Part A: general Vol. 25; no. 5; pp. 277 - 291
• Main Authors: Cascetta, Ennio, Cantarella, Giulio Erberto
• Format: Journal Article
• Language: English
• Published: Oxford Elsevier B.V 01.09.1991; Pergamon
• Subjects: Applied sciences; Exact sciences and technology; Ground, air and sea transportation, marine construction; Transportation planning, management and economics; Stochastic model; Traffic flow; Dynamic model; Traffic assignment; Traffic management; Transportation network
• DOI: 10.1016/0191-2607(91)90144-F
• ISSN: 0191-2607

In this article a doubly dynamis assignment model for a general network is presented. It is assumed that users' choices are based on information about travel times and generalized transportation costs occurred in a finite number of previous days and, possibly, in previous periods of the same day. The information may be supplied and managed by an informative system. In this context, path and link flows vary for different subperiods of the same day (within-day dynamics) and for different days (day-to-day dynamics). The proposed model follows a nonequilibrium approach in which both within-day and day-to-day flow fluctuations are modelled as a stochastic process. A model of dynamic network loading for computing within-day variable arc flows from path flows is also presented. The model deals explicitly with queuing at oversaturated intersections and can be formulated as a fixed point problem. A solution scheme for the doubly dynamic assignment model is presented embedding a solution algorithm for the fixed-point problem.