An agent-based simulation framework for analysing fall risk among older adults in the evacuation procedures

摘要

Poor performance during an emergency evacuation from a residential building can lead to serious injuries (even death) and loss of property. As a result, it is paramount that evacuation plans are designed to account for both the physical (or geometrical) limitations of the space, such as the stairway designs, and, more importantly, the physical condition of those living in the building. Generally, when egress planning is considered or modelled during the design phase of a building, it is assumed that its residents are sampled from an average population with an average mobility condition. As a result, older adults tend to be ignored or regarded as part of the average population in the evacuation simulations. Due to their vulnerability, it is common to see older adults suffer from falls, which may result in serious trauma, such as fractures, from which older adults often do not fully recover. Moreover, many of the adverse effects that can occur during an evacuation, including the risk of falling, is compounded when a larger number of individuals try to exit a building within a short time interval given the geometric constraints of the built environment and the (limited) physical capacity of the residents. To analyse the risk of falling among older adults in evacuation procedures, this paper proposes an agent-based simulation framework, which consists of three primary components: agent-based modelling, fall risk index establishment, and fall risk assessment for older adults in an evacuation scenario. A case study of an apartment-style seniors' residence in Edmonton, Canada is provided to demonstrate the applicability of the framework in evaluating fall risk among older adults. Furthermore, a sensitivity analysis is conducted to better understand how the number of evacuees and different designs of the built environment influence both the fall risk of individuals and the fall risk attributable to the built environment.