A Monte Carlo enhanced agent based model for Tokenized BIM Twin (TBT) resource reallocation in construction projects

Abstract Construction projects involve a complex relationship of interdependence among numerous parties, which often leads to delays and significant coordination costs among contractors. While approaches based on Building Information Modeling (BIM) and Digital Twin (DT) technologies have enhanced monitoring and visualization capabilities, they remain limited in their ability to enable autonomous, decentralized, and tamper-proof resource reallocation. This paper proposes and computationally evaluates a three-layer Tokenized BIM-Twin (TBT) framework under controlled simulation conditions. The framework combines a static IFC-based BIM layer for resource tokenization, a real-time Digital Twin layer to detect anomalies and measure deviations, and a blockchain-based smart contract layer that has the potential to minimize disputes and reallocate resources automatically and transparently under the modelled conditions. The framework is evaluated using an agent-based simulation model defined on an eight-parameter uncertainty space. A total of 150 Monte Carlo iterations per scenario ( n = 450 observations) are performed across three configurations: Baseline (traditional management), Centralized BIM + DT, and the TBT framework. The results are assessed using one-way ANOVA, Welch’s t-tests, and effect size analysis (Cohen’s d), indicating statistically significant differences across all evaluated performance dimensions under the modelled conditions. Schedule overrun is reduced within the simulation environment from a mean of 38.90% (± 6.24) to 3.30% (± 1.38) (F = 2672.5, p < 0.001, d = 7.88). Resource idle rates decline from 35.52% (± 3.22) to 20.04% (± 3.02) (d = 4.96), while disruption response time is reduced from 5.45 days (± 1.23) to 0.44 days (± 0.17) (d = 5.69). In addition to its conceptual contribution, the framework has practical implications, as the framework demonstrates the potential to improve coordination efficiency, reduce disputes, and increase transparency under the modelled conditions. The results suggest that the approach has potential for projects characterized by high uncertainty and complex stakeholder interactions, although empirical validation is still required before real-world performance can be confirmed.