Toward a Decision-Oriented Ontology of Resilience for Complex Engineered Systems

Resilience is routinely invoked to justify infrastructure investments and emergency-management strategies, yet it remains difficult to operationalize as an interpretable measurand for decision support in interdependent lifeline systems exposed to compounding hazards. A recurring source of confusion is the divide between engineered and ecological resilience. We argue that, in decision settings, this divide is often sustained by a category mistake: regimes and equilibria are used primarily as structured descriptions of acceptable operation rather than as ontologically thick system states. Building on ResiliOnt, we present a decision-oriented ontology sketch that extends the core resilience pattern with a decision-context layer, a regime-and-trajectory layer, and an assessment-and-mapping layer. Regime descriptions are modeled as Information Content Entities, which allows engineered resilience to be represented as the special case in which the acceptable regime set is singleton, while ecological-style resilience permits multiple acceptable regime descriptions and admissible transitions among them. We provide a compact competency-question traceability table, a minimal formal fragment, and a worked micro-scenario for Colorado Front Range lifelines. The example includes an explicit regime-description template, regime-membership rule, regime-shift criterion, resilience metric specification and value, evidence and provenance records, uncertainty annotation, and a miniature mapping to Bayesian and dynamic Bayesian network variables. The result is a more inspectable and reusable ontology pattern for resilience-informed decision support that clarifies measurand semantics, supports interpretability of metric outputs, and enables transparent interfaces to probabilistic causal models without embedding probabilistic semantics into the ontology itself.

Keywords: Resilience ontology, decision-oriented resilience, Information Content Entities, resilience metrics, probabilistic decision support, critical infrastructure.