A multiclass simulation-based dynamic traffic assignment model for mixed traffic flow of connected and autonomous vehicles and human-driven vehicles

• Published in: Transportmetrica (Abingdon, Oxfordshire, UK) Vol. 21; no. 2
• Main Authors: Bamdad Mehrabani, Behzad, Erdmann, Jakob, Sgambi, Luca, Seyedabrishami, Seyedehsan, Snelder, Maaike
• Format: Journal Article
• Language: English
• Published: Taylor & Francis 04.05.2025
• Subjects: Connected and Autonomous Vehicles (CAVs); Human Driven Vehicles (HDVs); mixed traffic flow; multiclass traffic assignment; Simulation-based traffic assignment
• ISSN: 2324-9935; 2324-9943
• DOI: 10.1080/23249935.2023.2257805

Connected and Autonomous Vehicles (CAVs) may exhibit different driving and route choice behaviours compared to Human-Driven Vehicles (HDVs), which can result in a mixed traffic flow with multiple classes of route choice behaviour. Therefore, it is necessary to solve the Multiclass Traffic Assignment Problem (TAP) for mixed traffic flow. However, most existing studies have relied on analytical solutions. Furthermore, simulation-based methods have not fully considered all of CAVs' potential capabilities. This study presents an open-source solution framework for the multiclass simulation-based TAP in mixed traffic of CAVs and HDVs. The proposed model assumes that CAVs follow system optimal with rerouting capabilities, while HDVs follow user equilibrium. It also considers the impact of CAVs on road capacity at both micro and meso scales. The proposed model is demonstrated through three case studies. This study provides a valuable tool that can consider several assumptions for better understanding the impact of CAVs on mixed traffic flow.