Modeling and Calibration of a Mixed Traffic Road Section in VISSIM with Multiple Measure of Effectiveness Through Genetic Algorithm

This chapter delves into the complexities of modeling and calibrating traffic simulation models in mixed traffic conditions, a challenge particularly prevalent in developing countries. The study focuses on an urban road section near Silchar, India, where traffic is characterized by a mix of vehicle types and poor lane discipline. The authors employ VISSIM, a widely-used traffic simulation software, to model the study area and use a genetic algorithm (GA) to calibrate the model. The GA optimizes sensitive driving behavior parameters, ensuring that the simulated traffic conditions closely match real-world observations. The chapter highlights the use of multiple measures of effectiveness (MoEs), including vehicular speed, acceleration/deceleration, and travel time, to validate the model. Additionally, the study examines the distribution of minimum headway and CO2 emissions, providing a holistic approach to traffic assessment. The methodology presented offers a significant advancement in traffic simulation, addressing the heterogeneity and complexity of real-world traffic conditions. The results demonstrate the effectiveness of using GA for model calibration, achieving high accuracy and reliability in traffic assessments.