Multi-Agent Debate System Based on Large Language Models: Structured Deliberation and Validation in Satellite Communications

Multi-Agent Debate System Based on Large Language Models: Structured Deliberation and Validation in Satellite Communications | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Multi-Agent Debate System Based on Large Language Models: Structured Deliberation and Validation in Satellite Communications Susana Gómez, Alejandro Mozo, Tomás Navarro, Sergio Gálvez, Francisco L. Valverde This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9575030/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 7 You are reading this latest preprint version Abstract Structured multi-agent debates among Large Language Models (LLMs) have emerged as a powerful paradigm for enhancing reasoning reliability and argumentative coherence. Motivated by the European Space Agency's (ESA) interest in trustworthy AI for space operations, this study proposes a moderated, domain-adaptive multi-agent debate framework applied to the high-stakes domain of satellite communications (SatCom). Specifically, it assesses (i) the efficacy of structured deliberation against single-agent baselines, and (ii) the impact of model heterogeneity versus homogeneity. A single-agent baseline is compared against a multi-agent framework deploying a moderator and two domain-specialized experts. These systems utilize local 70B-parameter LLMs in homogeneous (Llama-3.3) and heterogeneous (Llama-3.3 + DeepSeek-R1 + Qwen-2.5) configurations, all augmented with a shared, curated Retrieval-Augmented Generation (RAG) corpus combining academic institutional sources and ESA material from the Nebula portal (SatNex V programme). Outputs from 213 technical queries are evaluated via LLM-as-a-judge across three phases: baseline proficiency, strategic reasoning, and executive readiness. Single-agent systems lead in encyclopedic tasks, where retrieval suffices over deliberation. However, both multi-agent configurations outperform in strategic reasoning, with heterogeneous debates achieving superior performance in executive scenarios by victory margins of up to 2.75 points on a 10-point scale. These results validate architectural diversity as a decisive factor in resolving high-complexity technical conflicts. Ultimately, this work delivers a generalizable, fully traceable deliberation framework suitable for real-world, mission-critical environments. Code, prompts, and evaluation data are publicly available at https://github.com/amozo-es/multi-agent-debate/ . Large Language Models llm multi-agent satellite communications Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 05 May, 2026 Reviewers agreed at journal 04 May, 2026 Reviewers agreed at journal 04 May, 2026 Reviewers invited by journal 03 May, 2026 Editor assigned by journal 01 May, 2026 Submission checks completed at journal 01 May, 2026 First submitted to journal 30 Apr, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9575030","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":637237022,"identity":"a2ed0a5f-a824-4ee5-96bc-ccc4067a2377","order_by":0,"name":"Susana Gómez","email":"","orcid":"","institution":"University of Malaga","correspondingAuthor":false,"prefix":"","firstName":"Susana","middleName":"","lastName":"Gómez","suffix":""},{"id":637237023,"identity":"33679d7a-be9f-4e4f-ae17-01f000750728","order_by":1,"name":"Alejandro Mozo","email":"","orcid":"","institution":"University of Malaga","correspondingAuthor":false,"prefix":"","firstName":"Alejandro","middleName":"","lastName":"Mozo","suffix":""},{"id":637237024,"identity":"d77e2acc-51b1-43b7-8c7d-e18da70b83e8","order_by":2,"name":"Tomás Navarro","email":"","orcid":"","institution":"European Space Agency","correspondingAuthor":false,"prefix":"","firstName":"Tomás","middleName":"","lastName":"Navarro","suffix":""},{"id":637237025,"identity":"68d98041-89ad-4571-900b-44c195f9ec73","order_by":3,"name":"Sergio Gálvez","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAArElEQVRIiWNgGAWjYFAC5gYgYUOSFkaQljTStRwmQYO8+8E2iQ9/zieubT/A+PAHMVoMzyS2Sc7guZ247UwCszEPUVoaEpuNeSSAWg4ksEkT5TDD/ofNxn8MziVuO/+A/SdRDpOXSGx8zJBwIHHbjQQ2BqIcZiDxsPFhz4Fk4203HjZLE6VFvj/5wIEff+xkt51PPviRKIcZHIAzwRFEjC1EqhsFo2AUjIKRDADAEThOjyJoJgAAAABJRU5ErkJggg==","orcid":"","institution":"University of Malaga","correspondingAuthor":true,"prefix":"","firstName":"Sergio","middleName":"","lastName":"Gálvez","suffix":""},{"id":637237026,"identity":"3bd5c68b-7191-464c-a4ef-365f2a9f0cd9","order_by":4,"name":"Francisco L. Valverde","email":"","orcid":"","institution":"University of Malaga","correspondingAuthor":false,"prefix":"","firstName":"Francisco","middleName":"L.","lastName":"Valverde","suffix":""}],"badges":[],"createdAt":"2026-04-30 09:38:17","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9575030/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9575030/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":109068166,"identity":"4eed3390-9906-4883-a1f8-2900225f3318","added_by":"auto","created_at":"2026-05-12 10:04:12","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":965614,"visible":true,"origin":"","legend":"","description":"","filename":"MultiagentdebatesystembasedonLLMSpringer.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9575030/v1_covered_c1afa5f0-d5ba-4904-aa34-69b7d2d3f12a.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Multi-Agent Debate System Based on Large Language Models: Structured Deliberation and Validation in Satellite Communications","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"journal-of-intelligent-information-systems","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"jiis","sideBox":"Learn more about [Journal of Intelligent Information Systems](http://link.springer.com/journal/10844)","snPcode":"10844","submissionUrl":"https://submission.nature.com/new-submission/10844/3","title":"Journal of Intelligent Information Systems","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Large Language Models, llm, multi-agent, satellite communications","lastPublishedDoi":"10.21203/rs.3.rs-9575030/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9575030/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eStructured multi-agent debates among Large Language Models (LLMs) have emerged as a powerful paradigm for enhancing reasoning reliability and argumentative coherence. Motivated by the European Space Agency's (ESA) interest in trustworthy AI for space operations, this study proposes a moderated, domain-adaptive multi-agent debate framework applied to the high-stakes domain of satellite communications (SatCom). Specifically, it assesses (i) the efficacy of structured deliberation against single-agent baselines, and (ii) the impact of model heterogeneity versus homogeneity. A single-agent baseline is compared against a multi-agent framework deploying a moderator and two domain-specialized experts. These systems utilize local 70B-parameter LLMs in homogeneous (Llama-3.3) and heterogeneous (Llama-3.3 + DeepSeek-R1 + Qwen-2.5) configurations, all augmented with a shared, curated Retrieval-Augmented Generation (RAG) corpus combining academic institutional sources and ESA material from the Nebula portal (SatNex V programme). Outputs from 213 technical queries are evaluated via LLM-as-a-judge across three phases: baseline proficiency, strategic reasoning, and executive readiness. Single-agent systems lead in encyclopedic tasks, where retrieval suffices over deliberation. However, both multi-agent configurations outperform in strategic reasoning, with heterogeneous debates achieving superior performance in executive scenarios by victory margins of up to 2.75 points on a 10-point scale. These results validate architectural diversity as a decisive factor in resolving high-complexity technical conflicts. Ultimately, this work delivers a generalizable, fully traceable deliberation framework suitable for real-world, mission-critical environments. Code, prompts, and evaluation data are publicly available at \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/amozo-es/multi-agent-debate/\u003c/span\u003e\u003cspan address=\"https://github.com/amozo-es/multi-agent-debate/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/p\u003e","manuscriptTitle":"Multi-Agent Debate System Based on Large Language Models: Structured Deliberation and Validation in Satellite Communications","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-11 22:17:49","doi":"10.21203/rs.3.rs-9575030/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-05-05T04:01:31+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"43278928116016964553008349014797671961","date":"2026-05-05T02:34:52+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"322544554507678187575740175320817102508","date":"2026-05-04T14:41:29+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-05-04T00:44:06+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-05-01T14:19:27+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-05-01T12:39:39+00:00","index":"","fulltext":""},{"type":"submitted","content":"Journal of Intelligent Information Systems","date":"2026-04-30T09:22:36+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"journal-of-intelligent-information-systems","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"jiis","sideBox":"Learn more about [Journal of Intelligent Information Systems](http://link.springer.com/journal/10844)","snPcode":"10844","submissionUrl":"https://submission.nature.com/new-submission/10844/3","title":"Journal of Intelligent Information Systems","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"1bb8c055-aa4b-4904-8668-0421911ea397","owner":[],"postedDate":"May 11th, 2026","published":true,"recentEditorialEvents":[{"type":"editorInvitedReview","content":"","date":"2026-05-05T04:01:31+00:00","index":22,"fulltext":""},{"type":"reviewerAgreed","content":"43278928116016964553008349014797671961","date":"2026-05-05T02:34:52+00:00","index":21,"fulltext":""},{"type":"reviewerAgreed","content":"322544554507678187575740175320817102508","date":"2026-05-04T14:41:29+00:00","index":20,"fulltext":""},{"type":"reviewersInvited","content":"23","date":"2026-05-04T00:44:06+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-05-01T14:19:27+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-05-01T12:39:39+00:00","index":"","fulltext":""},{"type":"submitted","content":"Journal of Intelligent Information Systems","date":"2026-04-30T09:22:36+00:00","index":"","fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-05-11T22:17:49+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-11 22:17:49","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9575030","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9575030","identity":"rs-9575030","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}