An agent-based modeling approach for understanding the effect of worker-management interactions on construction workers' safety-related behaviors

Pervasive unsafe behaviors of construction workers are the primary cause of accidents on construction job sites. The workers' safety-related behaviors are subject to a variety of factors, such as interactions with coworkers and interventions by management teams. The impacts of these factors have attracted considerable attention in academia but are yet to be fully examined. To provide a quantitative assessment of these impacts and their managerial implications in practice, an agent-based modeling approach of construction safety-related behaviors, which adopts a bottom-up architecture, is proposed in this research. This approach regards safety performance as an emergent property of the behaviors and interactions of construction workers and management teams. The development of the agent-based model is based on a range of theoretical and empirical evidence, including a cognitive analysis model for modeling worker behaviors, and site observations and surveys for the design of simulated working environment and tasks, as well as the modeling of individual behaviors and interactions. Four managerial scenarios are simulated using the proposed approach. Based on the simulation results, the effects of a number of managerial factors on the safety performance of construction workers are examined. These factors include duties of supervisors, the competency and inspection strategy of safety officers, the frequency of safety training, senior managers' involvement in safety activities and emphasis on safety goals. The effects of these factors are examined in a quantitative manner, and relevant implications for construction safety management practice are discussed. The findings prove that agent-based modeling is an effective approach for analyzing the characteristics and patterns of construction safety-related behaviors and assessing possible safety management strategies.