Agent-Based Simulation of Construction Workflows Using a Relational Data Model

Ling Ma1 & Rafael Sacks2

1Postdoctoral Researcher, National Building Research Institute, Technion, Haifa, Israel, [email protected]
2Assoc. Prof., Faculty of Civil and Env. Eng., Technion, Haifa, Israel, [email protected]


To what extent is uncertainty concerning process status a cause of waste in construction workflows? Work studies and action research are expensive methods for investigation of such questions concerning construction workflow control policies and their results have limited applicability. Agent-based simulation (ABS) is particularly suitable for modelling peoples’ behavior and interaction in complex settings, like in construction, and therefore represents an alternative. We present a parametric ABS system (EPIC 2.0) developed using a relational data model for modelling construction workflow; the model enables users to specify the construction subjects (subcontractor trade crews), their work methods, the amount of work, the workspaces (locations), dependencies between the works, etc. The simulation encapsulates both variability and uncertainty in the construction workflow. Variability arising from design changes, quality checks and working conditions may lead to random change in workload and performance. Uncertainty arises from the fact that agents do not have full or perfect information. The major advantages of this ABS system are its ability to run differently configured virtual projects in terms of work crews, locations and production system control policies and to test the relative impacts of various approaches to communication of process status information. Simulation results conclude information asymmetry causes erroneous task maturity judgments and inappropriate work assignments, and of course affects the construction workflow.


Agent-based simulation; construction workflow; uncertainty; relational data model



Ma, L. & Sacks, R. 2016, 'Agent-Based Simulation of Construction Workflows Using a Relational Data Model' In:, 24th Annual Conference of the International Group for Lean Construction. Boston, Massachusetts, USA, 20-22 Jul 2016.

Download: BibTeX | RIS Format