Improving the effectiveness of fire prevention using the “premonition” agent-based model of domestic fire risk behaviours

Abstract

Purpose Understanding and predicting the behaviours of households within a community is a key concern for fire services as they plan to deliver effective and efficient public services. In this paper, an agent-based modelling approach is used to deepen understandings of changing patterns of behaviour within a community. The paper aims to discuss this issue. Design/methodology/approach This “Premonition” model draws on historical data of fire incidents and community interventions (e.g. home safety checks, fire safety campaigns, etc.) collated by South Yorkshire Fire and Rescue, UK, to unpack patterns of changing household behaviours within the region. Findings Findings from simulations carried out using the Premonition model, show that by targeting close-knit groups of connected households, the effectiveness of preventative interventions and utilisation of associated resources is enhanced. Furthermore, by repeating these interventions with the same households over time, risk factors within the wider area are further reduced. Originality/value The study thus shows that annual repeat visits to fewer and more targeted high-risk postcodes increase the overall reduction in risk within an area, when compared with a scattered coverage approach using one-off (i.e. not repeat) household visits within a postcode.