This document formalizes the Dynamic Threat Mitigation (DTM) model as a closed-loop control system under adversarial pressure. It defines states, observations, decisions, actuations, and evidence, and shows how CART and PQE alter system dynamics.

• S
  State: Asset posture, configs, crypto primitives, exposure graph.
• O
  Observations: Telemetry, detections, CART outcomes, partner signals.
• Π
  Policy: Versioned constraints and objectives (risk, performance, compliance).
• D
  Decision: Allow/deny/route, promote/rollback, cipher negotiation, exception grants.
• A
  Actuation: Gateway changes, mesh/mTLS policy, PKI updates, code-signing, key rotation.
• E
  Evidence: Signed artifacts linking O → D → A with timings and ownership.

• Minimize expected loss under attack given business impact and SLOs.
• Maintain crypto-agility and interoperability during PQC transition.
• Ensure explainability and auditability of every high-impact decision.

CART yields attack path likelihood and blast-radius projections. DTM weights these with data sensitivity and external exposure to schedule changes and allocate defenses.

• Guardrail constraints for performance and availability during changes.
• Rollback invariants (safe states) and bounded exception lifetimes.
• Graceful degradation and negotiated fallbacks to preserve connectivity.

• Risk: Probability of successful objective × impact.
• Efficacy: Delta in CART success rate after control changes.
• Resilience: MTTR, rollback success, steady-state variance of KPIs.