{"schema_version":1,"research_id":"eoa_inverse_innovation_exp05_external_evaluation_20260803","source_assessment_id":"invariant_mode_decomposition_design__information_theory:P5:v0","cell_id":"invariant_mode_decomposition_design__information_theory","search_queries":["delay tolerant network routing spectral radius transient matrix trapping mode routing optimization","opportunistic network routing Markov chain spectral analysis delivery probability relay nodes","site:ietf.org RFC delay tolerant networking Bundle Protocol store and forward routing","site:ccsds.org schedule-aware bundle routing standard CCSDS","\"spectral radius\" \"delay tolerant\" routing","\"absorbing Markov chain\" \"delay tolerant network\" routing","\"Markov chain\" \"routing policy\" spectral radius packet delivery","\"trapping\" modes routing network Markov chain packets","site:nasa.gov delay tolerant networking routing intermittent connectivity operational need Artemis LunaNet","site:nasa.gov DTN store and forward routing contact graph NASA ION","site:nasa.gov Delay Tolerant Networking handbook routing operational challenges","site:ccsds.org Schedule-Aware Bundle Routing CCSDS 734.3-B-1 pdf","Jain Fall Patra Routing in a Delay Tolerant Network SIGCOMM 2004 pdf","MaxProp routing vehicle-based disruption-tolerant networks paper pdf primary","RFC 6693 PRoPHET probabilistic routing protocol intermittently connected networks","Tassiulas Ephremides backpressure routing stability 1992 pdf","Boyd Diaconis Xiao fastest mixing Markov chain transition probabilities second largest eigenvalue modulus pdf","fastest mixing Markov chain graph spectral radius transition matrix optimization primary paper","\"minimize\" \"spectral radius\" transient matrix absorbing Markov chain","\"absorbing Markov chain\" optimize transition probabilities spectral radius","\"fastest absorption\" Markov chain graph transition probabilities","routing minimize spectral radius substochastic matrix packet","site:ojs.aaai.org \"Designing Fast Absorbing Markov Chains\"","\"Designing Fast Absorbing Markov Chains\" Ermon Gomes Sabharwal 2014","site:umass.edu MaxProp Routing Vehicle-Based Disruption-Tolerant Networks pdf Burgess Gallagher Levine","site:cs.umass.edu MaxProp pdf Burgess Gallagher Jensen Levine"],"sources":[{"source_id":"S1","title":"DTN Resources for Mission Developers","publisher":"NASA","url":"https://www.nasa.gov/technology/space-comms/delay-disruption-tolerant-networking-mission-resources/","source_class":"OFFICIAL_ORGANIZATION_DATA","publication_date":"2023-09-20","accessed_at":"2026-08-03","claims_supported":["NASA identifies delay, disconnection, data-rate mismatch, bandwidth/contact-time efficiency, and reliable delivery as DTN needs.","NASA reports operational DTN use on the ISS and other missions and identifies ION routing, testing, verification, and management capabilities.","NASA, mission operators, and infrastructure providers are credible adopters and authorizers for an offline routing-analysis study."]},{"source_id":"S2","title":"CCSDS All Active Publications and Delay Tolerant Networking Working Group","publisher":"Consultative Committee for Space Data Systems","url":"https://ccsds.org/publications/allpubs/","source_class":"STANDARD","publication_date":"2019-07","accessed_at":"2026-08-03","claims_supported":["CCSDS 734.3-B-1 specifies Schedule-Aware Bundle Routing for stable-topology, time-varying scheduled connectivity.","CCSDS identifies space-mission organizations and mission-support infrastructure providers as stakeholders and develops standards in response to operational needs.","The CCSDS DTN work addresses large delays, intermittent connectivity, and store-and-forward routing."]},{"source_id":"S3","title":"RFC 9171: Bundle Protocol Version 7","publisher":"Internet Engineering Task Force","url":"https://www.ietf.org/ietf-ftp/rfc/rfc9171.pdf","source_class":"STANDARD","publication_date":"2022-01","accessed_at":"2026-08-03","claims_supported":["BPv7 supports store-carry-forward communication under intermittent, scheduled, predicted, and opportunistic connectivity.","RFC 9171 deliberately leaves route computation and routing-information-base population outside the Bundle Protocol specification, leaving room for competing routing methods.","DTN congestion, authentication, confidentiality, bundle expiration, and at-rest protections constrain any operational routing modification."]},{"source_id":"S4","title":"RFC 6693: Probabilistic Routing Protocol for Intermittently Connected Networks","publisher":"Internet Research Task Force","url":"https://datatracker.ietf.org/doc/html/rfc6693","source_class":"OFFICIAL_GUIDANCE","publication_date":"2012-08","accessed_at":"2026-08-03","claims_supported":["PRoPHET is an established probabilistic DTN-routing analogue based on encounter history and transitivity.","Intermittent routing must trade delivery performance against scarce bandwidth, buffer space, and transmission resources.","PRoPHET requires sufficiently persistent contact circumstances to estimate predictability and documents routing attacks and buffer-overflow risks."]},{"source_id":"S5","title":"Routing in a Delay Tolerant Network","publisher":"ACM SIGCOMM","url":"https://conferences.sigcomm.org/sigcomm/2004/papers/p299-jain111111.pdf","source_class":"PRIMARY_RESEARCH","publication_date":"2004-08-30","accessed_at":"2026-08-03","claims_supported":["DTN routing is a constrained optimization problem over time-varying connectivity, finite buffers, bandwidth, and storage.","The paper explicitly reports that locally random forwarding can oscillate indefinitely among a node set or reach a dead end.","It supplies shortest-path, expected-delay, queue-aware, and linear-programming comparators and identifies the performance-versus-network-knowledge tradeoff."]},{"source_id":"S6","title":"MaxProp: Routing for Vehicle-Based Disruption-Tolerant Networks","publisher":"IEEE INFOCOM / University of Massachusetts Amherst","url":"https://groups.cs.umass.edu/jensen/wp-content/uploads/sites/17/2022/03/jensen-et-al-infocom2006.pdf","source_class":"PRIMARY_RESEARCH","publication_date":"2006-04","accessed_at":"2026-08-03","claims_supported":["MaxProp is a deployed DTN-routing analogue using historical path likelihoods, acknowledgments, previous-intermediary lists, and transmission/drop priorities.","The routing problem visibly involves limited contact duration and storage.","MaxProp was evaluated using 60 days of traces from a real 30-bus DTN, demonstrating feasibility of trace-replay comparison."]},{"source_id":"S7","title":"Designing Fast Absorbing Markov Chains","publisher":"Association for the Advancement of Artificial Intelligence","url":"https://ai.stanford.edu/~ermon/papers/aaai14-mcmc.pdf","source_class":"PRIMARY_RESEARCH","publication_date":"2014-07-27","accessed_at":"2026-08-03","claims_supported":["Prior research already formulates selection of transition probabilities in an absorbing Markov chain to minimize expected time to a sink.","Expected absorption time and its gradient can be computed from the transient matrix without simulating the underlying chain.","This substantially overlaps the candidate's transient-operator and forwarding-probability optimization core, although it does not provide the proposed DTN-specific modal interpretation and governance package."]},{"source_id":"S8","title":"Stability Properties of Constrained Queueing Systems and Scheduling Policies for Maximum Throughput in Multihop Radio Networks","publisher":"IEEE Transactions on Automatic Control / University of Maryland repository","url":"https://drum.lib.umd.edu/items/571fda52-aefb-4497-9a2d-69d8c7c907b9","source_class":"PRIMARY_RESEARCH","publication_date":"1992-12","accessed_at":"2026-08-03","claims_supported":["Backpressure's foundational prior art optimizes scheduling in constrained multihop packet networks for throughput stability.","Queue-stability-oriented routing is a strong comparator because it directly acts on congestion rather than inferred packet-location modes."]}],"problem_evidence":{"support":"STRONG","rationale":"Official NASA, CCSDS, and IETF materials establish that intermittent connectivity, long delays, limited contact opportunities, storage, congestion, and reliable delivery are consequential operational problems. Primary DTN research specifically observes that forwarding can oscillate indefinitely among a group of nodes even when local forwarding remains possible. What is not externally established is the prevalence of a reproducible, well-conditioned slow eigenmode as the cause of deadline failures in an operational single-copy relay network.","source_ids":["S1","S2","S3","S4","S5","S6"]},"stakeholder_evidence":{"support":"STRONG","rationale":"NASA is an identifiable adopter and potential funder: it reports operational DTN use, distributes ION and testing resources, and identifies missions and infrastructure operators that need reliable delivery and efficient contact-time use. CCSDS identifies mission organizations and mission-support infrastructure providers as stakeholders and publishes an operational routing standard. No named mission has expressed demand specifically for modal trapping analysis, so pull is strong for the parent routing problem but only inferred for this intervention.","source_ids":["S1","S2"]},"prior_art":{"proximity":"SUBSTANTIAL_COLLISION","closest_analogues":[{"name":"Designing Fast Absorbing Markov Chains","similarity":"It uses a transient matrix with sink states and selects transition probabilities to reduce expected absorption time—the candidate's central mathematical control problem.","remaining_difference":"The candidate adds DTN-specific packet-location eigenmodes, node participation maps, mode-conditioned sensitivity, residual and drift governance, and matched duty-cycle constraints. Whether those additions improve decisions over direct absorption-time optimization is untested.","source_ids":["S7"]},{"name":"Routing in a Delay Tolerant Network","similarity":"It formulates DTN routing as constrained optimization over time-varying contacts, buffers, delays, and capacity; it also identifies indefinite oscillation among nodes and supplies direct optimization baselines.","remaining_difference":"It does not choose controls by decomposing a fitted transient relay operator or preserve a governed slow-subspace explanation.","source_ids":["S5"]},{"name":"PRoPHET","similarity":"It learns predictable cross-node contact structure and redirects forwarding toward relays with greater delivery predictability while limiting bandwidth and buffer use.","remaining_difference":"It uses encounter-history/transitivity scores rather than eigenmodes of an end-to-end transient packet-location operator.","source_ids":["S4"]},{"name":"MaxProp","similarity":"It uses historical path likelihoods, loop-avoidance information, acknowledgments, and packet priorities to improve delivery under limited contacts and storage.","remaining_difference":"It does not expose or explicitly damp a consequential invariant packet-location subspace.","source_ids":["S6"]},{"name":"Schedule-Aware Bundle Routing and ION","similarity":"They are standardized or operational DTN routing practices for scheduled contacts and provide an adoption and implementation baseline.","remaining_difference":"They compute routes from contact schedules rather than diagnosing and damping empirically fitted slow packet-location modes.","source_ids":["S1","S2"]},{"name":"Backpressure scheduling","similarity":"It redirects packet service using network-wide queue consequences under constrained multihop capacity and is a strong rival for congestion-induced delay.","remaining_difference":"It targets queue stability and throughput, not persistence or oscillation of an undelivered-location probability mode.","source_ids":["S8"]}],"distinctive_claim_remaining":"For a frozen, approximately stationary, single-copy intermittent relay policy, a routing change selected to damp a reproducible and well-conditioned slow or oscillatory subspace of the transient packet-location operator will improve held-out deadline delivery under matched transmission, storage, priority, and duty-cycle budgets beyond the unchanged policy and beyond direct expected-absorption-time optimization, PRoPHET/MaxProp-style routing, schedule-aware/expected-delay routing, and backpressure—while its modal diagnostics provide earlier and more stable warning than link or queue metrics. This is contrastive and falsifiable but remains empirically unverified.","confidence":"HIGH"},"implementation_evidence":{"support":"MODERATE","rationale":"The computation is feasible for a frozen finite-state single-copy model: transient matrices, absorption-time objectives, gradients, and constrained routing optimization are established, while NASA provides DTN implementations and testing interfaces and real trace-driven studies demonstrate an offline workflow. Feasibility falls sharply when traffic interactions, replication, acknowledgments, priorities, correlated contacts, or rapid regime changes violate the Markov/stationarity assumptions. No source demonstrates this full modal workflow in a live DTN, validates eigenvector conditioning, or establishes authority for operational route changes.","source_ids":["S1","S3","S4","S5","S6","S7"]},"scores":{"meaningful_impact":{"score":4,"rationale":"Missed deadlines, wasted contact time, storage pressure, and energy/duty-cycle consumption matter in operational DTNs, including space missions and constrained mobile networks.","source_ids":["S1","S3","S4","S5","S6"]},"stakeholder_pull":{"score":4,"rationale":"NASA and CCSDS identify operational DTN adopters, standards, implementations, and explicit needs for reliable delivery and efficient contact use; modality-specific demand has not been expressed.","source_ids":["S1","S2"]},"incremental_advantage":{"score":2,"rationale":"The proposal may add interpretability and early-warning value, but direct absorption-time optimization already controls the same transient-chain transition probabilities, and mature DTN rivals directly optimize delivery, delay, resources, or stability.","source_ids":["S4","S5","S6","S7","S8"]},"distinctiveness_plausibility":{"score":3,"rationale":"The DTN-specific combination of slow-subspace diagnosis, node loadings, governed sensitivity, residuals, and drift checks is distinguishable, but its mathematical optimization core substantially collides with absorbing-chain design.","source_ids":["S5","S7"]},"technical_implementability":{"score":3,"rationale":"Offline estimation and eigensensitivity analysis are implementable with existing DTN traces and tools, but stationarity, non-normal conditioning, multi-packet interactions, and reliable estimation of Q are unresolved.","source_ids":["S1","S4","S5","S6","S7"]},"adoption_authority_feasibility":{"score":3,"rationale":"NASA mission and infrastructure operators and CCSDS stakeholders are identifiable, and an offline shadow study needs no live-routing authority. Operational adoption would require mission-specific network, security, safety, and relay-operator approval not evidenced here.","source_ids":["S1","S2","S3"]},"evidence_readiness":{"score":2,"rationale":"Methods, implementations, standards, comparators, and trace-replay precedents exist, but no candidate-specific trace, fitted operator, slow-mode prevalence estimate, or comparative result is available.","source_ids":["S1","S4","S5","S6","S7"]},"safety_net_benefit":{"score":4,"rationale":"The proposed first step is offline and reversible, while residual, drift, resource, security, and rollback checks could reduce the risk of harmful routing changes. These safeguards still need operational validation.","source_ids":["S1","S3","S4"]},"scalability":{"score":3,"rationale":"Finite-matrix analysis is computationally plausible for bounded relay networks, and DTN implementations span multiple mission contexts. Scaling the state to traffic class, queue, replication, custody, and contact regimes may cause state explosion and frequent refitting.","source_ids":["S1","S5","S7"]}},"score_confidence":"MODERATE","costs":{"first_evidence":{"band_2026_usd":"50K_TO_250K","scope":"A six-to-twelve-week offline study for one frozen topology and traffic class: trace preparation, Q estimation, conditioning and modal analysis, implementation of comparators, held-out replay, and a safety/interpretation report.","confidence":"MODERATE","assumptions":["Existing contact traces, packet logs, simulator, and baseline routing implementation are available without new hardware procurement.","Approximately one network researcher, one software/data engineer, and fractional operations/security review are required.","The estimate is a resource-equivalent range, not a vendor quote."],"source_ids":["S1","S5","S6","S7"]},"initial_deployment_startup":{"band_2026_usd":"250K_TO_1M","scope":"Integration into a shadow-routing or isolated test-network environment, telemetry and rollback instrumentation, configuration signing, scenario expansion, verification, and mission/operator review.","confidence":"LOW","assumptions":["No flight-hardware redesign or new relay is required.","Existing ION/BPv7 test interfaces and network emulation infrastructure can be reused.","Mission assurance and security review are bounded to a non-live canary."],"source_ids":["S1","S3"]},"operational_launch":{"band_2026_usd":"1M_TO_5M","scope":"Mission-grade implementation, independent verification and validation, security and safety assurance, protected-traffic testing, operator training, configuration governance, staged canary, and rollback qualification across affected nodes.","confidence":"LOW","assumptions":["Launch covers one network or mission family rather than a fleet-wide standard change.","No new spacecraft, gateway, spectrum allocation, or major hardware procurement is included.","Class-B or comparable mission-software assurance and multi-organization coordination are required."],"source_ids":["S1","S2","S3"]},"annual_recurring":{"band_2026_usd":"250K_TO_1M","scope":"Continuous telemetry ingestion, mode and residual monitoring, periodic refitting and regression testing, incident review, configuration maintenance, and operator/security support for one operational network.","confidence":"LOW","assumptions":["Existing network telemetry and operations staff are available.","One to three technical FTE equivalents plus compute, storage, and periodic assurance reviews are sufficient.","Major topology expansion or hardware replacement is excluded."],"source_ids":["S1","S3","S4"]}},"verified_pipeline_gates":{"externally_supported_problem":{"status":"YES","reason":"Official and primary sources establish intermittent-connectivity routing, constrained contacts, storage, delay, and even multi-node oscillation as real problems.","source_ids":["S1","S2","S3","S4","S5","S6"]},"externally_credible_adopter_or_authorizer":{"status":"YES","reason":"NASA is an operational DTN adopter with mission-development resources and implementations; CCSDS identifies mission and infrastructure organizations that authorize standards and operations.","source_ids":["S1","S2"]},"distinct_testable_incremental_claim":{"status":"YES","reason":"The remaining claim is whether modal targeting adds held-out delivery and warning value beyond direct absorption-time optimization and established DTN/queue-based rivals under matched resource constraints.","source_ids":["S4","S5","S6","S7","S8"]},"bounded_next_evidence_step":{"status":"YES","reason":"A single frozen topology, fixed policy, prespecified fit/holdout traces, bounded routing modification, explicit comparators, metrics, and falsifiers define an offline experiment.","source_ids":["S1","S5","S6","S7"]},"no_unresolved_safety_or_authority_stop":{"status":"YES","reason":"The next step is offline and changes no live routing, custody, contact schedule, priority, or duty cycle. Live deployment remains unauthorized until mission-specific security, safety, and operations approvals are obtained.","source_ids":["S1","S2","S3"]},"credible_cost_scope_and_range":{"status":"UNCERTAIN","reason":"The work packages are bounded and broad resource-equivalent bands are plausible, but no direct labor estimate, vendor quote, mission assurance budget, trace-access cost, or integration baseline was found.","source_ids":["S1","S3"]}},"next_evidence_step":"Run a six-to-twelve-week partnered offline replay on exactly one frozen topology, gateway, single-copy custody rule, traffic class, epoch definition, routing table, transmission budget, priority policy, and duty-cycle envelope. Predeclare at least 30 fitting and 30 held-out contact sequences or nonoverlapping trace blocks, with identical packet injections across policies. Fit Q only on the fitting partition; report bootstrap uncertainty, eigenvector condition numbers, slow-subspace angles, spectral separation, and residual structure. Select one bounded modal repair without observing holdout outcomes. On holdout data compare: unchanged routing; the modal repair; direct constrained expected-absorption-time optimization; SABR/CGR or expected-delay routing appropriate to the contact regime; PRoPHET or MaxProp; and backpressure, all with matched transmission, storage, priority, and duty-cycle budgets. The primary endpoint is deadline-delivery fraction; secondary endpoints are mean and 95th-percentile gateway hitting time, transmissions per delivered packet, peak node residence/storage, duty-cycle violations, mode drift, and structured residuals. Falsify the proposal if no reproducible consequential slow subspace exists; if the selected subspace is ill-conditioned or unstable across resamples; if coordinate-level causes explain the misses; if any safety/resource constraint is breached; or if the modal repair fails to exceed the strongest budget-matched rival on the predeclared primary endpoint by at least five percentage points without worsening protected secondary criteria. A positive result may authorize only a reversible shadow or isolated-network canary.","blocking_evidence":["No external evidence establishes how frequently real single-copy intermittent relay policies exhibit a reproducible slow or oscillatory packet-location subspace after accounting for single bottlenecks, source timing, destination outages, queues, acknowledgments, and traffic interactions.","No candidate-specific trace demonstrates that Q is stationary enough, well-conditioned enough, and sufficiently separated for node loadings and sensitivities to be reliable.","No held-out comparison shows incremental benefit over direct constrained absorption-time optimization, which substantially collides with the candidate's mathematical core.","No comparison against PRoPHET, MaxProp, schedule-aware or expected-delay routing, and backpressure has been performed under identical transmission, storage, priority, and duty-cycle budgets.","Actual mission-specific routing authority, protected-traffic rules, security review, rollback latency, and acceptable failure thresholds remain unidentified.","Cost ranges lack direct labor, vendor, assurance, integration, and data-access benchmarks."],"research_disposition":"PARTNERED_RESEARCH_PROGRAM","world_novelty_boundary":"The search found substantial mathematical prior art for designing absorbing Markov chains by changing transition probabilities to reduce sink-reaching time, plus extensive DTN routing standards and practices. It did not establish whether the exact DTN-specific package of slow-mode diagnosis, participation mapping, sensitivity-selected repair, residual governance, and drift withdrawal has previously been implemented. World novelty, patentability, freedom to operate, market size, and realized impact remain unmeasured.","arm":"COMPLETE_PROPOSAL_PORTFOLIO","candidate_version":0,"controller_recommendation":{"action":"STOP_EMPIRICAL_RESEARCH_NEEDED","repairable":false,"material_progress_observed":false,"progress_targets":["Obtain one authorized, de-identified operational or high-fidelity DTN trace set with packet locations, contacts, queues, priorities, transmissions, acknowledgments, storage, and duty-cycle data.","Demonstrate a reproducible, consequential, well-conditioned slow subspace on fitting data and stability on held-out contact sequences.","Run the predeclared budget-matched comparison against direct absorption-time optimization, unchanged routing, SABR/CGR or expected-delay routing, PRoPHET/MaxProp, and backpressure.","Show that any delivery gain is not explained by increased transmissions, storage, energy, duty cycle, degraded protected traffic, or removal of resilience paths.","Document mission-specific authorizers, security and safety constraints, rollback timing, and signed configuration governance before any live canary.","Replace resource-equivalent cost bands with partner-validated labor, integration, assurance, compute, and recurring-operations estimates."],"reason":"Bounded web research verifies the parent problem, credible adopters, implementation ingredients, and substantial prior-art collision. The remaining value proposition is an empirical claim about mode prevalence, stability, interpretability, and comparative routing performance that cannot be resolved by further web search; it requires trace access and controlled replay or live-system testing."},"proposal_index":5}