An Agent-based Simulation Approach to Urban Morphology Analysis

The analysis of urban morphology and its impact on ventilation is a critical task in urban planning. Accurate urban-scale computational fluid dynamics models (CFD) are sometimes too expensive for rough estimations. Effective and convenient computer tools that could depict the pattern of the urban roughness are in great need. This paper presents a simulation approach which is grounded in the agent-based modeling (ABM) paradigm. A multi-agent simulation system is implemented which treats air particles as intelligent individuals that avoid obstacles and look for outlets based on 3-D building database. Path finding algorithms are developed to locate potential air paths and assess the ventilation condition. Various parametric models are built for a qualitative analysis of the model. A preliminary quantitative comparison with data obtained from a CFD model is also carried out in real world context. The comparison shows reasonable agreement between the present method and the traditional CFD method, suggesting that the ABM simulation approach can provide new perspectives to the wind engineering domain.