Adaptive Orchestration Management System

Reconstruction-Governed Execution at Effectuation Time

The Adaptive Orchestration Management System (AOMS) is an execution-governance architecture for determining whether a previously authorized action remains eligible when execution is actually attempted.

AOMS separates two conditions that conventional systems frequently collapse:

Historical authorization != present execution permission

Historical authorization is evidence. Execution permission is a current determination reconstructed from authority-relevant conditions at or near effectuation time.

Governing Pipeline

AuthorityArtifact
    -> AuthorityState
    -> ContinuityVector
    -> BoundaryAssessment
    -> ReconciliationRecord
    -> EligibilityRecord
    -> ExecutionDecision

The canonical decision space is:

• ALLOW
• DENY
• ESCALATE
• REAUTHORIZE

An affirmative decision is unavailable unless the complete ordered pipeline positively establishes present execution eligibility.

Current Repository Baseline

Version 1 contains:

• 100 deterministic execution cases;
• 100 corresponding evidence artifacts;
• 10 state-family reports;
• a replayable Python runtime;
• doctrine and execution-specification documents;
• authority, policy, temporal, identity, dependency, resource, environmental, agentic, evidence, and compound-state coverage.

All 100 v1 cases are deliberately inadmissible controls. They demonstrate that changed execution conditions defeat automatic execution despite prior authorization.

Version 1 is retained as a historical baseline. Phase II will implement the complete canonical pipeline and a balanced four-decision verification corpus.

Existing State Families

Range	State family
AOMS-001 - AOMS-010	Authority
AOMS-011 - AOMS-020	Policy
AOMS-021 - AOMS-030	Temporal
AOMS-031 - AOMS-040	Identity
AOMS-041 - AOMS-050	Dependency
AOMS-051 - AOMS-060	Resource
AOMS-061 - AOMS-070	Environment
AOMS-071 - AOMS-080	Agentic
AOMS-081 - AOMS-090	Evidence
AOMS-091 - AOMS-100	Compound

Run the Existing Demonstration

From the repository root:

python .\src\main.py --case .\cases\AOMS-001.json
python .\src\main.py --all

The v1 runtime produces deterministic violation findings and writes evidence artifacts under reports/json.

Phase II

Phase II will add:

• separate reconstruction, continuity, boundary, reconciliation, eligibility, and decision engines;
• canonical immutable state objects;
• explicit governance and execution contexts;
• authority and delegation reconstruction;
• agent-readable boundary evaluation;
• complete provenance chains;
• four-outcome decision semantics;
• failure-injection and bypass testing;
• balanced positive and negative controls;
• multi-agent and cross-boundary execution cases;
• conformance specifications and formal proof obligations.

See architecture/PHASE_II_ARCHITECTURE_MANIFEST.md and ROADMAP.md.

Foundational Publication

Ashley S. Harris, "Adaptive Orchestration Management System (AOMS): Reconstruction-Governed Execution Eligibility Determination for Autonomous and Distributed Systems," 2026.

DOI: https://doi.org/10.5281/zenodo.20673754

Software Record

DOI: https://doi.org/10.5281/zenodo.20819505

Author

Ashley S. Harris Independent Researcher ORYNTH Systems

Status

• v1 corpus: complete historical baseline
• Phase II canonicalization: active
• Phase II runtime: pending Batch 02
• Phase II verification corpus: pending Batch 03
• specifications and manuals: pending Batch 04
• publication extensions: pending Batch 05