Dynamic scenario-based DEMATEL for modeling risk topology in autonomous maritime systems

Abstract Autonomous maritime systems operate in dynamically changing environments where risk structures are not invariant but depend on operational regimes. Traditional decision-making trial and evaluation laboratory (DEMATEL) applications assume a fixed causal topology and therefore may underestimate structural transformations that occur during transitions between normal, degraded, and emergency states. This study proposes a dynamic scenario-based extension of the DEMATEL method for modeling time-evolving risk topology in autonomous maritime systems. The framework introduces scenario-dependent total-relation matrices, a transformation operator for quantifying structural shifts, a scenario instability index to measure regime sensitivity of factors, and an integral dynamic criticality index for identifying strategically significant elements of the system. The proposed approach reinterprets risk topology as a state-dependent structure and enables formal analysis of structural inversion phenomena, where dominant causal drivers change across operating modes. The results demonstrate that static causal mapping may overlook regime-dependent amplification effects, while the dynamic formulation provides deeper insight into structural resilience and adaptive safety management. The proposed framework extends DEMATEL from static influence analysis to regime-dependent structural modeling. A scenario-based case study and sensitivity analysis confirm the presence of structural inversion effects and the numerical stability of the proposed approach.