Learning-Based Topology Generalized Planning for Distribution Networks with Soft Open Points

Learning-Based Topology Generalized Planning for Distribution Networks with Soft Open Points | 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 Article Learning-Based Topology Generalized Planning for Distribution Networks with Soft Open Points Jiawei Zhou, Linkang Zhou, Yong Wang, Chaoping Xu, Yin Ren This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9197078/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 5 You are reading this latest preprint version Abstract The increasing structural variability of modern distribution networks, driven by distributed energy resource integration, reconfiguration, and expansion, challenges conventional topology-dependent planning methods, particularly for soft open point deployment, where repeated re-optimization under topology changes leads to high computational cost and limited scalability. This paper proposes a learning-based topology-generalized planning framework that enables transferable decision-making across heterogeneous network configurations by modeling the system as a graph and learning topology-invariant embeddings that capture structural and operational characteristics. Based on these embeddings, a neural policy network directly generates siting and sizing decisions, while physical feasibility is ensured through the integration of power flow constraints and device limits via differentiable optimization layers and penalty-based formulations. A scenario-based training strategy with domain randomization and meta-learning enhances robustness and adaptability to unseen topologies. Extensive experiments on more than 1,200 network topologies and 60,000 scenario instances demonstrate that the proposed method achieves a normalized optimality gap of 2.8%–3.6% across both seen and unseen systems, significantly outperforming heuristic and graph-agnostic baselines that exceed 9%–12% under topology shifts, while maintaining stable performance under uncertainty with cost increases limited to approximately 17% compared to over 60% for deterministic models, and reducing inference time to around 0.02 seconds per instance. The proposed framework provides a scalable and robust solution for distribution network planning by enabling topology-invariant decision-making with physical consistency. Physical sciences/Engineering Physical sciences/Mathematics and computing Physical sciences/Physics Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviewers invited by journal 21 Apr, 2026 Editor assigned by journal 19 Apr, 2026 Editor invited by journal 17 Apr, 2026 Submission checks completed at journal 11 Apr, 2026 First submitted to journal 11 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-9197078","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":627755686,"identity":"d416cea3-a7c2-4055-bce1-0db405ca7743","order_by":0,"name":"Jiawei Zhou","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA0klEQVRIiWNgGAWjYBACxgYGNiAlIcfPzHzgwIcfxGuxMJZsb0s8OLOHOItAWioSN5w5Y3yYg40I9cwz0p89+NgmkbjhRs6Hwww8DPL8YgcIOGxGQrrhzDYJ45k3cjccLrBgMJw5O4GAltkJx6R52yRk+0BaZvAwJBjcJqglsQ2khbHhRs6DwzxsRGlJZgNpUZxw5gwDkVrmP2OTnHFOAhTIBsBAliDsF8Oe488kPpTVgaLy8YcPP2zk+aUJaWlA5UvgVw4C8oSVjIJRMApGwYgHAIkJSBnVc9/uAAAAAElFTkSuQmCC","orcid":"","institution":"Suzhou Electric Power Design Institute Co.,Ltd","correspondingAuthor":true,"prefix":"","firstName":"Jiawei","middleName":"","lastName":"Zhou","suffix":""},{"id":627755687,"identity":"7e61a0cb-d43a-45d0-8732-3729a59855cc","order_by":1,"name":"Linkang Zhou","email":"","orcid":"","institution":"Suzhou Electric Power Design Institute Co.,Ltd","correspondingAuthor":false,"prefix":"","firstName":"Linkang","middleName":"","lastName":"Zhou","suffix":""},{"id":627755688,"identity":"95c4051b-414c-4d91-a1c3-8384f8c4f189","order_by":2,"name":"Yong Wang","email":"","orcid":"","institution":"Suzhou Electric Power Design Institute Co.,Ltd","correspondingAuthor":false,"prefix":"","firstName":"Yong","middleName":"","lastName":"Wang","suffix":""},{"id":627755690,"identity":"4cb5d179-11f0-4e6d-ab00-dd98d8824f4e","order_by":3,"name":"Chaoping Xu","email":"","orcid":"","institution":"Suzhou Electric Power Design Institute Co.,Ltd","correspondingAuthor":false,"prefix":"","firstName":"Chaoping","middleName":"","lastName":"Xu","suffix":""},{"id":627755691,"identity":"1309f425-6570-479c-916d-5b3f93656e68","order_by":4,"name":"Yin Ren","email":"","orcid":"","institution":"Suzhou Electric Power Design Institute Co.,Ltd","correspondingAuthor":false,"prefix":"","firstName":"Yin","middleName":"","lastName":"Ren","suffix":""}],"badges":[],"createdAt":"2026-03-23 07:41:30","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9197078/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9197078/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":107870668,"identity":"c41f1b5f-7263-4dac-b15f-595fdaf01c50","added_by":"auto","created_at":"2026-04-27 07:40:20","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":654077,"visible":true,"origin":"","legend":"","description":"","filename":"LearningBasedTopologyGeneralizedPlanningforDistributionNetworkswithSoftOpenPoints1.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9197078/v1_covered_a89a7077-5416-4a4e-b500-f7561f6f7f2f.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Learning-Based Topology Generalized Planning for Distribution Networks with Soft Open Points","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":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-9197078/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9197078/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"The increasing structural variability of modern distribution networks, driven by distributed energy resource integration, reconfiguration, and expansion, challenges conventional topology-dependent planning methods, particularly for soft open point deployment, where repeated re-optimization under topology changes leads to high computational cost and limited scalability. This paper proposes a learning-based topology-generalized planning framework that enables transferable decision-making across heterogeneous network configurations by modeling the system as a graph and learning topology-invariant embeddings that capture structural and operational characteristics. Based on these embeddings, a neural policy network directly generates siting and sizing decisions, while physical feasibility is ensured through the integration of power flow constraints and device limits via differentiable optimization layers and penalty-based formulations. A scenario-based training strategy with domain randomization and meta-learning enhances robustness and adaptability to unseen topologies. Extensive experiments on more than 1,200 network topologies and 60,000 scenario instances demonstrate that the proposed method achieves a normalized optimality gap of 2.8%–3.6% across both seen and unseen systems, significantly outperforming heuristic and graph-agnostic baselines that exceed 9%–12% under topology shifts, while maintaining stable performance under uncertainty with cost increases limited to approximately 17% compared to over 60% for deterministic models, and reducing inference time to around 0.02 seconds per instance. The proposed framework provides a scalable and robust solution for distribution network planning by enabling topology-invariant decision-making with physical consistency.","manuscriptTitle":"Learning-Based Topology Generalized Planning for Distribution Networks with Soft Open Points","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-27 04:03:26","doi":"10.21203/rs.3.rs-9197078/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewersInvited","content":"","date":"2026-04-21T14:00:52+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-04-19T22:39:27+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-04-17T19:33:12+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-04-11T07:25:45+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2026-04-11T07:22:51+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"7907d0e7-efdf-476a-aae1-7ff4b1ae8e5d","owner":[],"postedDate":"April 27th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":66816734,"name":"Physical sciences/Engineering"},{"id":66816735,"name":"Physical sciences/Mathematics and computing"},{"id":66816736,"name":"Physical sciences/Physics"}],"tags":[],"updatedAt":"2026-04-27T04:03:26+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-27 04:03:26","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9197078","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9197078","identity":"rs-9197078","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}