Understanding post-pandemic work-from-home behaviours and community level energy reduction via agent-based modelling

An agent-based model (ABM) was developed to simulate post-pandemic work-from-home (WFH) behaviours based on social behaviour theories (conservation of resources theory and small-world theory). Key decision factors affecting WFH behaviour are identified, including environment constraints (social influences of family members and colleagues), resource constraints, and personal stress tolerance. The model is validated by a previous full-time worker survey which collected almost 2000 effective responses. Scenario simulations were carried out to understand the impacts of these factors on WFH behaviour. Three sets of results are generated: (1) the proportion of employees choosing to work at the office, (2) the proportion of employees choosing to WFH, and (3) community level energy consumption compared to business-as-usual. Analysis shows that environment and resource constraints discourage employees from WFH, leading to 8–12% more workers preferring the office as the workplace, so that community level energy saving is less significant (14–21% variation). However, as agents adapt to new a WFH arrangement and personal stress tolerance increases, more agents choose WFH for consecutive days, with 19% fewer workers choosing the office. The community level energy consumption saving increases by 42% when employees have higher stress tolerance. Analysis across all four simulated scenarios reveals that improving personal stress tolerance is the most effective means for achieving more significant community level energy reduction. The modelling results are compared with peer reports, and policy implications are discussed. This study contributes to the understanding of socio-environmental aspect of WFH by shedding light on worker decision-making processes when selecting their workplace based on behaviour theories and quantifying the variation of community level energy reduction according to different WFH arrangements.