Li-BIM, an agent-based approach to simulate occupant-building interaction from the Building-Information Modelling

Abstract Building design involves many challenges and requires to take into account the interaction between the building and the users. Different occupant behaviour models implemented with building simulation tools (thermal, air quality, lighting) have been proposed. Among these, models based on the agent approach seem to be the most promising. However, existing models poorly describe human cognition and the social dimension. Moreover, they are often oriented towards a specific use (thermal simulation, waste management) without being transposable to another field, and they require a significant instantiation effort for each new case, making their use difficult. This article proposes an agent-based model called Li-BIM that simulates the behaviour of the occupants in a building and their indoor comfort. Li-BIM model is structured around the numerical modelling of the building –BIM- (with standard exchange format IFC), a high-resolution cognitive model, and the coupling with various physical models. Li-BIM simulates the reactive, deliberative and social behaviour of occupants in residential dwellings based on the Belief–Desire–Intention architecture. This model, thanks its ease of use and flexibility, is an operational and relevant tool to support building design process with a human-centred approach. An application of the model is presented, focusing on energy consumption and the inhabitants’ comfort. In-situ data obtained from the instrumented house that served as case study have been compared with simulation results from Li-BIM and a standard energy simulation software, demonstrating the reliability of the proposed model.