Moshe Cohen

Analysis of the Breakdown Process on Congested Freeways

The aim of the study is to model the flow breakdown phenomenon on congested freeways. Because of the importance of the breakdown phenomenon, the present research focuses on analyzing, understanding, and modeling heavy-flow breakdowns when no apparent bottlenecks exist. It is hypothesized that breakdowns are caused by internal microscopic flow oscillations and the transition from dense flow to unstable flow on freeways. A new density-versus-time model was developed to describe this phenomenon. The model shows the density behavior over time, which increases sharply toward the peak period and then decreases and increases again toward the breakdown. This new revelation is described by a combination of an exponential model, a logistic model, and a weighting function. This density–time relationship is calibrated by using data collected from a congested freeway at 1-min intervals. The beginning of the unstable flow is defined as the drop in speed below 60 km/h; it is shown that breakdown is defined as a process that starts at the local minimum of the new model and ends at the beginning of the unstable flow. It was found that the model well describes the behavior of density over time, from the increase in density through the breakdown process and into the rapid density-fluctuation phase of the unstable flow.

Models for Estimating Drivers Following on Two-Lane Rural Highways

Traffic performance on two-lane rural highways is commonly characterized by percent time spent following (PTSF), a parameter that determines the level of service on a road segment. Although PTSF is frequently used by transportation engineers and researchers, there is no agreement about its estimation method and values. Therefore, a study was done to calibrate and compare models of PTSF with the data collected on 25 segments of 15 two-lane highways in Israel. The study showed large differences between the collected models and previous model values. Furthermore, the study estimated traffic intensity, which was an additional characteristic of the driver-following phenomenon; this variable did not approach its expected value of 1.0. This lack of agreement between actual and expected values may result from changing driver behavior, that is, impatience as platoon lengths and delays increase. To avoid long platoons, drivers prefer to travel between them and then to perform a passing maneuver sooner than expected. This phenomenon not only reduces platoon length and delays, but also increases the risk of serious and fatal crashes.

dEVElOpMEnT OF an inFRaSTRucTuRE cOEFFiciEnT BY an analYTic hiERaRchY pROcESS and iTS RElaTiOnShip TO SaFETY

This paper describes a study which had as its primary objectives to develop a numerical Infrastructure Coefficient (IC) representing the overall infrastructure characteristics of a highway and to develop a crash prediction model correlating the IC with crash rates on two-lane rural highways. The first phase of the study, involving the development of the IC, examined the correlation between different infrastructures characteristics and crash rate. The Analytic Hierarchy Process (AHP) was also used to develop the IC. Its function was to attribute a specific weight to each infrastructure characteristic in order to reflect its relative importance to road safety when compared with other characteristics considered in the study. Finally, a model predicting crash rates based on the proposed IC was calibrated and presented. The study concluded that a well-built and maintained highway could reduce crashes by an average of 44% as compared to a highway with poor infrastructure.

SimWiz - a Self-Learning Tool for Learning Discrete Event System Simulation

A new tool for self-learning of simulation modeling is described that is especially utilizable in e-learning. The tool is a computer program called SimWiz (Simulation Wizard) that leads the user through well-defined steps. In each step the user is requested to respond to a simple request until the model is prepared. The final model can be quite complicated in spite the fact that each step in itself is simple. The model is translated to a commercial simulation program, Arena, and by examining it the student learns the use of building blocks of Arena.

Methodology of Discrete System Simulation Modeling with SimWiz

A methodology is described that leads the builder of a discrete event simulation through all the aspects of the modeling. The methodology is formalized as a wizard based process where the user responds to specific requests partly based on responses to earlier answers of the user. The embodiment of the methodology is a computer “wizard” called SimWiz. SimWiz produces outputs that facilitate participation of parties who may contribute to the model but possibly are not familiar with simulation software.