Scopus

The translation of cybersecurity policy into enforced practice in cloud-native and operational-technology environments is not a one-time event but a continuous process, susceptible to silent divergence between the authored policy and the configuration effectively realised by the infrastructure. We call this divergence compliance drift. Existing tooling addresses individual stages of the policy-to-practice translation but does not close the loop with the deployed infrastructure in the temporal regime where drift occurs: periodic auditing operates with latencies of days to weeks, static policy-as-code verification stops at compile time, runtime engines evaluate individual requests in milliseconds without reference to the global policy, and LLM extraction is one-shot. The framework targets a closed loop operating at the twinning rate of the Sensing Layer, with sub-second granularity in cloud-native instrumentation. This paper examines the Digital Twin as a candidate artefact for closing this gap. Through a functional analysis of its canonical characteristics, the dual representation of physical and virtual entities, the synchronisation governed by twinning rate and metrology, the bidirectional channels between physical and virtual systems, and the cognitive operations hosted within the virtual processes, we show that the Digital Twin already possesses, structurally, the properties required to address the three failure modes of the translation pipeline. These structural enablers are articulated in a four-layer framework, the Cognitive Digital Twin, supported by four formal definitions and a Bounded-Drift Proposition that sustains a compliance-drift bound under four explicit antecedents on knowledge-graph coverage, sensing fidelity, translation coverage, and extraction fidelity. The contribution is theoretical and conceptual; empirical validation is identified as future work.

In transport and logistics systems, decision-making is increasingly influenced by uncertainty stemming from demand variability, technological disruptions, and systemic risks present in supply chains. In these contexts, organizations need approaches that are rooted in data and analysis to assess key elements affecting system resilience and performance. Although current studies widely utilize stochastic and fuzzy models for operational decision-making, there has been insufficient focus on the systematic assessment of human-centric system elements—especially competencies—as decision variables in intricate logistics systems. This research proposes an analytical framework for multi-criteria decision-making that is driven by data and aimed at evaluating the significance of various competencies that affect labor market competitiveness and the adaptability of supply chains. The approach combines expert assessment with statistical and information-theoretic metrics, utilizing Kendall’s coefficient of concordance for evaluating consistency, Shannon entropy for analyzing distributional uncertainty, and the Gini coefficient for measuring concentration. This integrated method allows for the measurement of both variability and inequality within decision frameworks in the face of uncertainty. The findings indicate that hands-on experience and professional skills play a crucial role in decision-making structures, whereas the ability to adapt to technological advancements and a commitment to ongoing learning greatly enhance system resilience. The entropy results reveal a significant degree of structural balance in the decision criteria, while the low Gini values affirm a lack of concentration, indicating a distributed and multi-dimensional decision-making environment. The study provides analytical insights into the structure and relative importance of competencies in decision-making contexts related to supply chain resilience.