{"schema_version":1,"assessment_id":"eoa_inverse_innovation_exp03_opportunity320_20260801","source_experiment_id":"eoa_inverse_innovation_exp03_full320_20260801","cell_id":"computability_boundary_mapping__operations_research","archetype_slug":"computability_boundary_mapping","domain_slug":"operations_research","title":"Enforced solvable model classes with status-preserving fallbacks","opportunity_summary":"Conditionally replace a universal exact-termination claim for executable optimization models with an enforceable decidable fragment and explicit UNKNOWN, timeout, rejection, or weaker-mode outcomes. The opportunity could prevent unsupported infeasibility decisions, but it depends on proving that deployed callback semantics can encode unbounded computation and that useful models remain covered.","adopter_authorizer":"The OR platform owner can authorize claim, validation, and routing changes; solver engineers implement them; model owners authorize model migration; operational decision owners authorize reliance on weaker outputs.","scores":{"meaningful_impact":{"score":4,"rationale":"If the stated failure exists, separating infeasibility from timeout or undecidability could prevent valid plans from being rejected and unsupported outputs from driving allocation or scheduling decisions. The affected volume and realized harm are not established."},"stakeholder_pull":{"score":2,"rationale":"The candidate identifies platform teams, model authors, planners, and downstream decision owners, but supplies no observed complaints, adoption requests, failure frequency, or other evidence that these stakeholders prioritize the problem."},"incremental_advantage":{"score":4,"rationale":"Unlike the timeout baseline or the heuristic-and-compute rival, the proposal directly changes the admissible input class and preserves UNKNOWN and timeout labels, addressing both the claimed computability boundary and false-infeasibility pathway. Advantage remains conditional on useful coverage."},"distinctiveness_plausibility":{"score":3,"rationale":"The combination of a reviewed reduction, constructive fragment, enforceable membership, and labeled routing is coherent and testable, but prior art is explicitly unsearched, so distinctiveness cannot be affirmed closed-book."},"technical_implementability":{"score":3,"rationale":"Formalizing one callback family and testing bounded fixtures appears feasible, while proving the boundary, constructing a total exact solver, enforcing membership against adversarial models, and preserving production status semantics may be difficult. Deployed-language details are absent."},"adoption_authority_feasibility":{"score":4,"rationale":"The candidate allocates authority among platform, model, and operational owners and provides an offline first step with rollback. Full adoption still requires coordination across model migration and downstream decision policies."},"evidence_readiness":{"score":4,"rationale":"The proposal specifies load-bearing premises, independent proof review, bounded fixtures, comparison against current routing, and explicit problem and intervention falsifiers. No actual semantic audit, proof, or fixture result is supplied."},"safety_net_benefit":{"score":4,"rationale":"Explicit UNKNOWN and timeout states, restrictions on irreversible use of weak outputs, and rollback conditions provide a concrete safety net against unsupported infeasibility decisions. Benefit fails if downstream systems coerce those labels."},"scalability":{"score":3,"rationale":"Fragment checking and status-preserving routing could become platform-wide controls, but each callback language and model family may require separate semantics, proofs, migration work, and coverage validation."}},"score_confidence":"MODERATE","costs":{"first_evidence":{"band_2026_usd":"50K_TO_250K","scope":"Offline study of one representative model family: formalize deployed callback semantics, attempt and independently review the reduction, define a candidate decidable fragment, and compare current versus restricted routing on bounded fixtures.","confidence":"MODERATE","assumptions":["A representative model family and audit fixtures are accessible.","The work uses a small team with formal-methods, solver, and platform expertise.","No production routing or operational decision is changed.","Independent proof review and evaluation labor are included."]},"initial_deployment_startup":{"band_2026_usd":"250K_TO_1M","scope":"Build and integrate a fragment-membership validator, restricted solver path, explicit status taxonomy, audit logging, and non-production or shadow routing for one service.","confidence":"LOW","assumptions":["One existing platform is modified rather than replaced.","A constructive solver for the selected fragment is available or can be adapted.","The callback interface and downstream status contracts are documented.","Security, compliance, testing, and initial model-owner coordination are included."]},"operational_launch":{"band_2026_usd":"1M_TO_5M","scope":"Production hardening, model inventory and migration, downstream integration changes, decision-policy approvals, monitoring, documentation, training, staged rollout, and evaluation across the initial platform.","confidence":"LOW","assumptions":["Multiple operational model families and consumers require compatibility work.","Coverage loss must be measured before routing restrictions are enforced.","Approximate or one-sided modes require governance safeguards.","The launch does not require replacing the entire optimization platform."]},"annual_recurring":{"band_2026_usd":"250K_TO_1M","scope":"Maintain fragment definitions and validators, review language and solver changes, monitor label integrity and coverage, investigate evasions, support model owners, and periodically reassess decision policies.","confidence":"LOW","assumptions":["The platform and accepted model language continue evolving.","Specialist engineering and formal review remain necessary.","Monitoring covers downstream coercion of UNKNOWN or timeout states.","Major platform replacement and compute-intensive universal solving are excluded."]}},"research_burden":"HIGH","earliest_credible_horizon":"12_TO_36_MONTHS","pipeline_gates":{"recognizable_externally_supportable_problem":{"status":"YES","reason":"The sealed candidate states a recognizable platform failure: unrestricted executable models enter an exact path, divergent runs are collapsed into infeasible or failure, and downstream systems may consume unsupported verdicts."},"identifiable_adopter_or_authorizer":{"status":"YES","reason":"Platform owners can change claims and routing, model owners can approve migration, and operational decision owners can govern reliance on weaker outputs."},"distinct_testable_incremental_claim":{"status":"YES","reason":"The proposal claims that enforceable solvable classes plus status-preserving fallback routing will reduce false infeasibility relative to timeout-based routing or additional heuristics and compute, subject to proof and coverage tests."},"bounded_next_evidence_step":{"status":"YES","reason":"The candidate authorizes an offline test on one model family with formal semantics, independent reduction review, bounded fixtures, a current-routing comparison, and explicit stop conditions."},"no_unresolved_safety_or_authority_stop":{"status":"YES","reason":"The first step is offline, retains production routing, forbids irreversible reliance on weak outputs, identifies relevant authorities, and stops on semantic mismatch, proof gaps, or status-label corruption."},"implementation_cost_scope_and_range":{"status":"UNCERTAIN","reason":"The candidate scopes the technical components and coordination parties but provides no staffing, model inventory, integration complexity, compliance requirements, or migration volume from which a reliable implementation range can be established."}},"blocking_evidence":["Evidence that deployed callback semantics actually permit faithful simulation of unbounded computation rather than only finite or forced-total models.","An independently validated reduction for the unrestricted deployed class and a constructive total-solver argument for the proposed fragment.","Evidence that fragment membership can be enforced, including against evasive or adversarial models.","Bounded comparison showing fewer false-infeasibility outcomes without unacceptable operational-model coverage loss.","Evidence that downstream systems preserve UNKNOWN, timeout, rejection, and weaker-mode labels rather than coercing them into infeasible.","Adoption evidence from platform, model, and operational owners, plus prior-art research sufficient to assess incremental distinctiveness."],"next_evidence_step":"Offline, select one representative accepted model family; formalize its deployed callback semantics; have an independent reviewer assess both a halting reduction for the unrestricted class and a constructive total-solver argument for a candidate fragment; then compare current timeout-based routing with fragment-checked, status-preserving routing on bounded fixtures. Falsify advancement if the semantics cannot encode unbounded computation, either proof fails, label integrity degrades, or false-infeasibility reduction requires unacceptable coverage loss.","research_questions":["Do deployed callbacks support unbounded computation, or are all accepted inputs already finite, bounded, or forced-total?","Can independent reviewers validate the unrestricted-class reduction and the constructive exact solver for the proposed fragment?","Can fragment membership be decided and enforced before exact routing, including for adversarial inputs?","How do false-infeasibility rates, termination, and useful-model coverage compare with the current timeout baseline on bounded fixtures?","Will downstream consumers preserve UNKNOWN and timeout semantics, and what policy changes are needed to prevent automatic irreversible actions?","Which platform, model, and operational owners would approve migration, and what evidence thresholds would they require?","Does external prior-art research show that the proposed composition offers a material incremental advantage over existing validation and solver-status practices?"],"recommendation":"PARTNERED_RESEARCH","uncertainty_constraints":["The candidate is a hypothesis and contains no observed performance or adoption results.","Problem prevalence, affected workload volume, stakeholder demand, and realized impact are unsupported.","The computability conclusion is conditional on the actual expressiveness of deployed callback semantics.","Decidability does not establish practical tractability or acceptable model coverage.","Prior art and world-level distinctiveness are unmeasured.","Cost bands are resource-equivalent planning ranges based only on the stated scope, not quotes or an audited system inventory.","The credible horizon assumes successful formal review, enforceable classification, and coordinated downstream migration."],"closed_book_prior_art_boundary":"Prior art is explicitly unsearched. This assessment makes no claim that computability-boundary proofs, decidable modeling fragments, solver-status taxonomies, or their proposed composition are novel, uncommon, or absent from existing systems."}