Integrating digital twin and agent-based simulation to support adaptive production system design in building projects

Production system design in building projects must address uncertainties such as fluctuating labor, supply disruptions, and variable site conditions to ensure efficiency and resilience. Static preconstruction planning alone is insufficient under these evolving constraints. This paper presents a decision-support framework for adaptive production system design, integrating an ontology-based digital twin with parametric agent-based simulation to enable real-time data ingestion, status model instantiation, and scenario analysis. The framework embeds Lean Construction principles, such as pull planning and continuous flow control, into its logic, allowing planners to proactively manage variability and enhance system responsiveness. By combining digital twins and agent-based simulation, the system supports the comparison of alternative production configurations based on projected performance. The framework's viability was demonstrated through qualitative, quantitative, and operational validations in two case study projects, highlighting its potential to support continuous adaptation in production system planning. • Introduced a DSS for adaptive production system design in building projects. • Combined Digital Twin and ABS for real-time simulation and decision support. • Enabled planners to test alternative plans with current project status data. • Reduced decision latency by automating data ingestion and model instantiation. • Validated in real projects through expert feedback and operational deployment.