An agent-based model for mixed traffic in Vietnam based on virtual local lanes

Abstract

The rapid urbanization and urban spread in developing countries create a huge number of challenges urban planners have to face. As an example, congestion has become a major issue in Vietnam and caused huge economic losses as well as an increase in air pollutant emissions. Tackling such an issue requires huge improvements in the city organization and infrastructure. Choosing the most appropriate solution among existing ones or imagining new ones require tools to understand their impact and assess their feasibility, effectiveness and acceptability. Computer simulation is a tool of choice to assess the interactions between dynamics at both macro-level (the city) or micro-level (a single road or a crossroad). Existing tools have nevertheless mainly been designed for cities with traffic dominated by cars that are supposed to respect regulations. These assumptions cannot be applied to traffic in Vietnam which is characterized by a mix of cars, motorbikes and other vehicle types and low respect of road lanes. This paper proposes an agent-based model simulating in a more realistic way the traffic in Vietnam using the GAMA platform. We illustrate its capabilities on two characteristic toy case studies (straight road and crossroad) and a real case study in rush hour and lower traffic situations.