Traffic Management System Performance Using Regression Analysis

This study presents a preliminary exploration in the use of regression analysis for evaluating long-run traffic management system performance. The following traffic management systems in the Twin Cities metropolitan area of Minneapolis were evaluated using multiple regression analysis: 1) ramp metering, 2) variable message signs, 3) Highway Helper Program, and 4) high occupancy vehicle (HOV) systems. Link speed and incident rate were separately used as the response variable. Although it is shown that regression analysis can be a simple and effective research method to test the macroscopic association between traffic management and traffic system performance, the authors conclude that additional research is needed to obtain overall evaluation of each of the traffic management systems. The authors also note that improvements are achievable through model improvement, adding relevant predictor variables, and decreasing limitations in data.