Ph.D Thesis
Ph.D Student Sofer Tal Development of Models for Congestion Analysis on Freeway Systems Department of Civil and Environmental Engineering Professor Abishai Polus Professor Shlomo Bekhor

Abstract

Congestion is a dynamic variable, including elements of space and time, which occurs on a daily basis. Since congestion involves high financial costs it is important to define and objectively compare different sites and different networks in respect to severity of their congestion. In addition, it is highly important to investigate what parameters influence the congestion, both local congestion in a given link and congestion in a freeway network system.

Congestion in a given link was defined according to a function describing observations of speed as a function of occupancy. This definition is mathematically rationalized and varies among different sites depending on the distance from the previous entrance. This definition is also sensitive to the lane in which it is being calculated in a freeway cut.

System congestion is highly dependent on the flexibility of the network for which it is been calculated. Since no prior study has dealt with the flexibility of a freeway network in regard to links that might have differing levels of congestion, a flexibility model was developed. The model defines the options that drivers have when choosing what route in the network to use for their trip. It is being based on perceived lengths of links the drives sense, which depend on the presence of congestion and traffic in the different links, number of available routes and how similar the routes are.

It was found that the flexibility measure is sensitive to the amount of congestion in the freeway system, and therefore, the model that is proposed is flow-dependent. It was also found that when the common lengths between the routes increase, the flexibility decreases, and that this knowledge can assist decision making regarding the number of effective routes that exists between a given origin - destination.

Since the state of traffic and congestion constantly changes throughout the day, the flexibility definition is dynamic.

The flexibility of a network with low traffic and no congestion in all links is then presented as a benchmark for the route flexibility with traffic in it. Any additional congestion that may occur can impact the flexibility negatively and therefore, the ratio between these two flexibilities is defined as the perceived system congestion. Having this definition, any change in the network can be calculated to determine how much this change helps decrease the perceived system congestion, and different networks can be compared.