Agent-based simulation of workers' behaviors, productivity, and safety around construction obstacles

Construction sites involve many hazardous areas and fixed and movable objects around which the workers interact. Unguided site policies or workers' behaviors can lead to productivity loss and (or) accidents. In this paper, an agent-based modeling and simulation framework has been developed to: (1) provide a detailed site model with temporary facilities and obstacles (danger zones and productivity-hindering spots); (2) model the workers' natural behaviors towards safety while moving around the site; and (3) quantify the consequent site productivity and accident potential of any site configuration. The paper discusses the site model and the behavioral rules that autonomous agents follow to simulate the movement of aggressive and avoider workers around hazard areas. The results of two case studies and sensitivity analyses show that different simulations with different site improvement measures can quantify the effectiveness of, and optimize, site operational measures and establish rewards for positive worker's behaviors that improve productivity and safety.