Expert-derived influence structure and structural associations among resilience engineering indicators in an Iranian petrochemical complex: an integrated fuzzy DEMATEL–MMDE–SEM study

Abstract Background Resilience engineering emerges from interactions among organizational and operational capabilities, yet these interactions are rarely mapped and empirically tested within a unified analytical framework. Methods This study integrated fuzzy decision-making trial and evaluation laboratory (DEMATEL), maximum mean de-entropy (MMDE), and covariance-based structural equation modeling to examine 12 resilience engineering indicators in an Iranian petrochemical complex. Expert judgments from 10 specialists were used to map the influence network, and survey data from 343 employees were used to test the MMDE-retained relationships. Results In the expert-derived DEMATEL network, management commitment, awareness, reporting culture, learning culture, teamwork, and buffering capacity occupied net-cause positions, whereas preparedness, flexibility, self-organization, redundancy, anticipation, and fault-tolerance occupied net-effect positions. MMDE retained 10 of 132 candidate relationships. The structural model showed acceptable fit (CFI = 0.959; RMSEA = 0.027), and all 10 prespecified paths were positive and significant. Awareness showed its strongest associations with self-organization and flexibility and had significant total effects on preparedness and fault-tolerance, partly through buffering capacity and flexibility. Conclusions These findings distinguish overall network position from path-specific influence and provide convergent evidence that awareness and usable adaptive capacity are central to resilient performance. Given the single-site, cross-sectional design, the identified directions should be interpreted as empirically supported structural associations rather than definitive causal effects.