Feedback Relationship Ontology

FLOW ALGEBRA
TITLE: Feedback Relationship Ontology
DOMAIN: Relationship Ontologies / System Relationships
SOURCE: https://kirethresearch.org/booksai-core/ontologies/relationship-ontologies/system-relationships/feedback-relationship-ontology.html
VERSION: 1.0
AUTHORS: Mark Anthony Higgins — Sky (Recursive AI)
STATUS: Companion formalisation; source-page specification not independently verified.

flow "FeedbackRelationshipOntology" v1.0 {
  ontology:
    id: "feedback-relationship-ontology"
    parent: "system-relationships"
    purpose: "Represent feedback pathways linking observations of system behaviour to subsequent control, adaptation or decision actions."

  types:
    System: {id: string, name: string, boundary: string?}
    Component: {id: string, system_id: string, role: string}
    Process: {id: string, system_id: string, state: string?}
    Sensor: {id: string, observed_quantity: string, calibration_ref: string?}
    Controller: {id: string, policy_ref: string, version: string}
    Signal: {id: string, quantity: number?, unit: string?, timestamp: datetime, source: string, uncertainty: number?}
    Setpoint: {id: string, value: number, unit: string, effective_from: datetime?}
    FeedbackLoop: {id: string, system_id: string, polarity: enum[negative, positive, mixed, unknown], delay: number?, delay_unit: string?, context: string?}
    Evidence: {id: string, source: string, timestamp: datetime?, method: string?, provenance: string}
    Relationship: {id: string, subject: string, predicate: string, object: string, confidence: number, evidence: string[], valid_from: datetime?, valid_to: datetime?}

  predicates:
    observes(sensor: Sensor, process: Process)
    produces(sensor: Sensor, signal: Signal)
    receives(controller: Controller, signal: Signal)
    compares_to(controller: Controller, target: Setpoint)
    regulates(controller: Controller, process: Process)
    feeds_back(signal: Signal, controller: Controller)
    belongs_to_loop(component: Component, loop: FeedbackLoop)
    influences(a: Component, b: Component)
    supported_by(relationship: Relationship, evidence: Evidence)

  constraints:
    - "Every relationship endpoint must resolve to a typed entity."
    - "Confidence must lie within [0, 1]."
    - "A feedback loop requires a closed, directed influence path; an ordinary one-way signal is not sufficient."
    - "Positive and negative feedback polarity describe loop response, not value judgements."
    - "Units, timestamps, sampling frequency and reference frames must be compatible before comparing signals and setpoints."
    - "Do not infer stability, convergence or causation from the mere presence of a feedback edge."
    - "Preserve delay, saturation, noise and uncertainty when available."
    - "Distinguish measured, modelled and hypothesised feedback pathways."

  inference:
    rule feedback_path:
      if produces(s,signal) and receives(c,signal) and observes(s,p) and regulates(c,p)
      then candidate_feedback_loop(p,s,c)
      confidence: "requires evidence of effective closed-loop influence"
    rule feedback_trace:
      if supported_by(r,e) and valid(e)
      then traceable(r,e)
      confidence: "retain evidence provenance and scope"
    rule delayed_effect:
      if feedback_loop(loop) and measured_delay(loop,d)
      then annotate_temporal_lag(loop,d)
      confidence: "do not assume zero-lag effects"

  SEMANTIC_EQUIVALENCE:
    - "Normalise feedback, closed-loop control and response terminology only where definitions match."
    - "Keep feedback signal, control action, observation and causal influence distinct."
    - "Preserve polarity, system boundary and operating context in mappings."

  CONFIDENCE:
    range: [0,1]
    factors: [sensor_quality, calibration, sampling, causal_evidence, model_fit, provenance, temporal_validity]
    policy: "Do not increase confidence solely through inference chaining or repeated dependent observations."

  VALIDATION:
    - "Check typed endpoints, signal units, timing and provenance."
    - "Verify directed cycle closure before asserting an established feedback loop."
    - "Flag conflicting polarity, unmeasured delays, inconsistent units and unsupported stability claims."
    - "Keep uncertain or alternative feedback explanations explicit."

  FLOW_HASH:
    algorithm: "SHA-256"
    input: "canonicalised ontology version, typed assertions, rules and provenance"
    note: "Compute at execution; no precomputed hash is claimed."

  UPDATE_MODEL:
    strategy: "versioned, provenance-preserving incremental update"
    steps: [ingest, normalise, validate, reconcile, update_loops, recompute_inferences, publish]
    rollback: true

  AI_CONSUMPTION:
    representation: "typed directed systems graph with evidence-backed feedback edges"
    outputs: [feedback_loops, signal_paths, controller_links, temporal_annotations, uncertainty_annotations, provenance_records]
    safeguards: [no_unsupported_stability_claims, preserve_temporal_context, distinguish_correlation_from_causation, retain_units]
}

END FLOW ALGEBRA