P49 — AIEP — Controlled Outlier Generation Engine (ts)

Applicant: Neil Grassby
Inventor: Neil Grassby
Classification: withheld — internal
Priority: Claims priority from GB2519711.2 filed 20 November 2025

Abstract

A controlled outlier generation engine for distributed arbitration systems operating under an Architected Instruction & Evidence Protocol (AIEP). A schema-defined divergence constraint set governs deterministic pointer-scoped mutation operations applied to a base artefact. Validated outliers are assembled into a canonical DivergenceMatrix and a DeterministicOutlierHash is computed over the ordered matrix. Non-permitted mutations cause fail-closed suppression of execution. The invention enables reproducible, constraint-bound divergence for robustness testing while preserving deterministic integrity and replay certification capability.

Field of the Invention

[0001] The present invention relates to deterministic divergence control in distributed computing systems.

[0002] More particularly, the invention relates to a controlled outlier generation engine configured to produce constraint-bound divergent artefacts within distributed arbitration systems operating under an Architected Instruction & Evidence Protocol (AIEP) system as defined in United Kingdom patent application GB2519711.2.

Background

[0003] Deterministic arbitration systems may require controlled exploration of alternative state configurations for development, robustness testing, or adversarial validation.

[0004] Unbounded divergence may compromise invariant compliance, replay certification, and distributed reproducibility.

[0005] Conventional outlier or mutation engines frequently rely upon stochastic processes or non-deterministic variation strategies.

[0006] Such approaches prevent reproducible enumeration of generated variants.

[0007] Known systems do not provide a deterministic divergence matrix governed by schema-bound constraints and hash-certifiable outlier identity.

[0008] There exists a need for an outlier generation mechanism that:

(a) produces bounded divergence;

(b) enforces invariant constraints during divergence;

(c) enumerates outliers deterministically; and

(d) generates reproducible outlier identity hashes.

Summary of the Invention

[0009] The invention provides a computer-implemented method for controlled outlier generation within a deterministic arbitration substrate.

[0010] A base artefact is selected for outlier generation.

[0011] A schema-defined DivergenceConstraintSet defines permissible mutation paths.

[0012] A deterministic pointer-scoped mutation engine enumerates candidate divergence operations.

[0013] Mutation operations are applied in canonical pointer order.

[0014] Each generated outlier artefact is validated against invariant constraints.

[0015] Non-compliant outliers are rejected.

[0016] Valid outliers are appended to a DivergenceMatrix in canonical order.

[0017] A DeterministicOutlierHash is computed over the ordered DivergenceMatrix.

[0018] Execution enablement is suppressed where constraint violations are detected.

[0019] The technical effect is reproducible, constraint-bound divergence enabling controlled exploration without compromising deterministic integrity.

Brief Description of the Drawing

FIG. 1 — Controlled Outlier Generation

   ┌────────────────────────────────────────────┐
   │   Primary Canonical Evidence Stream        │
   │   E1 ── E2 ── E3 ── E4 ── E5              │
   └─────────────────────────────────────────────┘
                    │
                    │ outlier generation engine
                    │ (bounded operator set)
   ┌────────────────▼────────────────────────────┐
   │   Structured Dissent (Outlier Ledger)       │
   │                                             │
   │   O1: parameter relaxation variant          │
   │   O2: context promotion variant             │
   │   O3: claim weakening variant               │
   │                                             │
   │   Each: hash-chained · deviation scored     │
   │         preserved · not elevated            │
   └─────────────────────────────────────────────┘

Definitions

[0020] DivergenceConstraintSet: A schema-defined set of permissible mutation operations and structural boundaries.

[0021] DivergenceMatrix: An ordered collection of validated outlier artefacts generated from a base artefact.

[0022] DeterministicOutlierHash: A cryptographic hash computed over the canonical ordered DivergenceMatrix.

[0023] PointerScopedMutation: A mutation operation applied to a specific CanonicalPointer path within a structured artefact.

Brief Description of the Drawings

Figure 1 illustrates base artefact selection and constraint binding.
Figure 2 illustrates deterministic pointer-scoped mutation enumeration.
Figure 3 illustrates construction of the DivergenceMatrix.
Figure 4 illustrates computation of DeterministicOutlierHash and fail-closed enforcement.

Detailed Description of Preferred Embodiments

Base Artefact Selection

[0024] A base artefact is selected from the genealogical structure.

[0025] Base artefact selection is deterministic and schema-bound.

[0026] Base artefact identity is incorporated into outlier generation records.

Divergence Constraint Definition

[0027] DivergenceConstraintSet is defined within a canonical schema.

[0028] DivergenceConstraintSet specifies permissible mutation types, structural limits, and quantitative thresholds.

[0029] Constraint definitions are version-pinned.

Deterministic Mutation Enumeration

[0030] PointerScopedMutation operations are enumerated in canonical pointer order.

[0031] Mutation operations comprise value substitution, bounded increment or decrement, structural insertion, or structural removal where permitted.

[0032] Enumeration order is deterministic and independent of runtime implementation.

[0033] No stochastic element is permitted.

Invariant Validation

[0034] Each generated outlier artefact is validated against invariant constraints.

[0035] Validation comprises schema compliance, lineage integrity, and structural correctness.

[0036] Outliers violating constraints are rejected and excluded from DivergenceMatrix.

Divergence Matrix Construction

[0037] Valid outlier artefacts are appended to DivergenceMatrix in canonical order.

[0038] DivergenceMatrix is serialised in canonical form.

[0039] DeterministicOutlierHash is computed over the canonical serialisation of DivergenceMatrix and base artefact identifier.

[0040] Nodes operating over identical base artefacts and constraint sets produce identical DivergenceMatrix and DeterministicOutlierHash.

Fail-Closed Enforcement

[0041] ExecutionEnablementSignal generation requires compliance with DivergenceConstraintSet.

[0042] Detection of non-permitted mutation or invariant violation results in fail-closed suppression.

[0043] No outlier bypasses invariant enforcement.

Certification Integration

[0044] An OutlierGenerationCertificate may be appended to the genealogical structure.

[0045] The certificate comprises base artefact identifier, DivergenceConstraintSet identifier, DeterministicOutlierHash, and schema version identifier.

[0046] OutlierGenerationCertificate enables replay-certifiable divergence reproduction.

Technical Effect

[0047] The invention enables reproducible outlier generation under deterministic constraints.

[0048] The invention prevents uncontrolled entropy injection into deterministic arbitration systems.

[0049] The invention produces hash-certifiable divergence matrices.

[0050] The invention improves robustness testing while preserving invariant governance.

[0051] The invention ensures identical divergence enumeration across distributed nodes.

Claims

1. A computer-implemented method for controlled outlier generation within a distributed arbitration system operating under an Architected Instruction & Evidence Protocol (AIEP), the method comprising:

   (a) selecting a base artefact;

   (b) applying a schema-defined DivergenceConstraintSet;

   (c) deterministically enumerating PointerScopedMutation operations in canonical pointer order;

   (d) generating candidate outlier artefacts;

   (e) validating candidate outlier artefacts against invariant constraints;

   (f) constructing an ordered DivergenceMatrix of validated outliers;

   (g) computing a DeterministicOutlierHash over the DivergenceMatrix; and

   (h) suppressing execution enablement in a fail-closed manner upon detection of non-permitted mutation.

2. The method of claim 1 wherein DivergenceConstraintSet defines permissible mutation types and quantitative thresholds.

3. The method of claim 1 wherein DeterministicOutlierHash is computed over canonical serialisation of DivergenceMatrix and base artefact identifier.

4. The method of claim 1 wherein mutation enumeration is independent of runtime implementation ordering.

5. The method of claim 1 further comprising generating an OutlierGenerationCertificate appended to a genealogical structure.

6. A distributed computing system configured to perform the method of any preceding claim.

7. A non-transitory computer-readable medium storing instructions which, when executed, perform the method of any of claims 1–5.