Agent-based simulation of trust development in human–robot teams: an empirically validated decision-support framework

Human–robot teams are increasingly deployed in complex operational environments, yet validated decision-support tools to predict team performance prior to deployment remain scarce. This paper presents an empirically grounded agent-based model (ABM) that captures trust dynamics, workload distribution, and collaborative performance in human–robot teams, intended as a decision-support instrument for practitioners selecting team configurations and parameter settings under resource constraints. The model, implemented in NetLogo 6.4.0, simulates teams of 2–10 agents performing tasks of varying complexity. We validate model behavior against Hancock et al.’s ( 2021 ) meta-analysis of human–robot interaction studies using two criteria: interval validity (whether simulated correlations fall within the published 95% confidence intervals) and ordinal validity (whether the simulation reproduces the relative ranking of effect magnitudes). The model achieved interval validity for half of the meta-analytic predictor categories and strong ordinal validity, indicating that the relative importance of trust antecedents is faithfully reproduced. Sensitivity analyses using one-factor-at-a-time and full factorial designs identified robot reliability as the dominant driver of trust, task success, and productivity, consistent with meta-analytic findings. Scenario analysis revealed that emergent cumulative trust asymmetry can fall below the per-event asymmetry configured in the model when trust-repair pathways remain active, identifying a boundary condition that extends existing theory. Scenario analysis also revealed substantial decoupling between trust magnitude and team performance, establishing calibration error—the discrepancy between subjective trust and objective capability—as a critical diagnostic distinct from trust magnitude itself. Factorial ANOVA confirmed significant main effects for reliability, transparency, communication, and collaboration, with minimal interaction effects, supporting an additive treatment of design levers in optimization formulations. The open-source implementation provides an evidence-based decision-support environment for exploring human–robot team configurations and is amenable to integration with formal optimization, design-of-experiments, and risk-analysis workflows familiar to operations research practitioners.