{"id":{"repo_id":"tu-berlin","oai_identifier":"oai:depositonce.tu-berlin.de:11303/25218"},"canonical_url":"https://search.dev.ndltd.org/etd/tu-berlin/oai:depositonce.tu-berlin.de:11303/25218","repository":{"repo_id":"tu-berlin","name":"Technische Universität Berlin","base_url":"https://api-depositonce.tu-berlin.de/server/oai/request"},"display":{"title":"Advancing mixed-integer programming using data-driven and deduction-based methods","abstract":"Mixed-Integer Problems (MIPs) form one of the most general classes of optimization problems. As they are used to model many real-world scenarios, solving MIPs efficiently is crucial. Most solvers are based on the well-known Branch-and-Bound algorithm, which utilizes different subroutines to help find an optimal solution faster. In this dissertation, we focus on two of the most impactful components of MIP solving: Primal heuristics and cutting planes. First, we present two data-driven learning frameworks, offline and online, that aim to optimize the use of heuristics by learning from data describing their behavior. These approaches are able to improve performance of an state-of-the-art open-source MIP solver on a broad set of homogeneous and heterogeneous instances. In the second part of this thesis, we derive strong cutting planes to enforce quadratic constraints present in a MIP. By applying monoidal strengthening, we strengthen intersection cuts by exploiting integrality information. In addition, we show that, in our setting, unique lifting exists, implying that our strengthening procedure leads to the strongest cut coefficients. Finally, we present a general framework for cut generation to identify conditions under which a family of cutting planes yield a polyhedral closure. This allows us to show polyhedrality for a broad range of popular cuts more easily.","abstract_html":"Mixed-Integer Problems (MIPs) form one of the most general classes of optimization problems. As they are used to model many real-world scenarios, solving MIPs efficiently is crucial. Most solvers are based on the well-known Branch-and-Bound algorithm, which utilizes different subroutines to help find an optimal solution faster. In this dissertation, we focus on two of the most impactful components of MIP solving: Primal heuristics and cutting planes. First, we present two data-driven learning frameworks, offline and online, that aim to optimize the use of heuristics by learning from data describing their behavior. These approaches are able to improve performance of an state-of-the-art open-source MIP solver on a broad set of homogeneous and heterogeneous instances. In the second part of this thesis, we derive strong cutting planes to enforce quadratic constraints present in a MIP. By applying monoidal strengthening, we strengthen intersection cuts by exploiting integrality information. In addition, we show that, in our setting, unique lifting exists, implying that our strengthening procedure leads to the strongest cut coefficients. Finally, we present a general framework for cut generation to identify conditions under which a family of cutting planes yield a polyhedral closure. This allows us to show polyhedrality for a broad range of popular cuts more easily.","abstract_has_math":false,"creators":["Chmiela, Antonia"],"institution":null,"degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":[],"advisors":["Pokutta, Sebastian"],"committee_chairs":[],"committee_members":[],"year":2025,"date_issued":"2025","date_published":"2025","updated_at":"2026-07-27T21:28:37Z","subjects":[],"languages":["en"],"rights":[],"rights_urls":["https://creativecommons.org/licenses/by/4.0/"],"identifier_entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://doi.org/10.14279/depositonce-24040"],"render_values":[{"text":"https://doi.org/10.14279/depositonce-24040","href":"https://doi.org/10.14279/depositonce-24040","code":true}]}]},"links":{"outbound_url":"https://depositonce.tu-berlin.de/handle/11303/25218","outbound_label":"Repository record","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Pokutta, Sebastian"]},{"key":"dc:creator","label":"Author","values":["Chmiela, Antonia"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2025-10-20T16:13:20Z"]},{"key":"dc:date.available","label":"Dc Date Available","values":["2025-10-20T16:13:20Z"]},{"key":"dc:date.issued","label":"Date","values":["2025"]},{"key":"dc:type","label":"Dc Type","values":["Doctoral Thesis"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language.iso","label":"Language (ISO)","values":["en"]},{"key":"dc:rights.uri","label":"Rights URI","values":["https://creativecommons.org/licenses/by/4.0/"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://depositonce.tu-berlin.de/handle/11303/25218","https://doi.org/10.14279/depositonce-24040"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Mixed-Integer Problems (MIPs) form one of the most general classes of optimization problems. As they are used to model many real-world scenarios, solving MIPs efficiently is crucial. Most solvers are based on the well-known Branch-and-Bound algorithm, which utilizes different subroutines to help find an optimal solution faster. In this dissertation, we focus on two of the most impactful components of MIP solving: Primal heuristics and cutting planes. First, we present two data-driven learning frameworks, offline and online, that aim to optimize the use of heuristics by learning from data describing their behavior. These approaches are able to improve performance of an state-of-the-art open-source MIP solver on a broad set of homogeneous and heterogeneous instances. In the second part of this thesis, we derive strong cutting planes to enforce quadratic constraints present in a MIP. By applying monoidal strengthening, we strengthen intersection cuts by exploiting integrality information. In addition, we show that, in our setting, unique lifting exists, implying that our strengthening procedure leads to the strongest cut coefficients. Finally, we present a general framework for cut generation to identify conditions under which a family of cutting planes yield a polyhedral closure. This allows us to show polyhedrality for a broad range of popular cuts more easily.","Gemischt-ganzzahlige Probleme (MIPs) bilden eine der allgemeinsten Klassen von Optimierungsproblemen. Da sie zur Modellierung vieler realer Szenarien verwendet werden, ist ihre effiziente Lösung von zentraler Bedeutung. Die meisten Lösungsverfahren basieren auf dem bekannten Branch-and-Bound-Algorithmus, der verschiedene Subroutinen nutzt, um schneller eine optimale Lösung zu finden. In dieser Dissertation konzentrieren wir uns auf zwei der wirkungsvollsten Komponenten: Primale Heuristiken und Schnittebenen. Zunächst stellen wir zwei datengetriebene Lernframeworks vor – eines im Offline- und eines im Online-Kontext – die darauf abzielen, den Einsatz von Heuristiken durch Lernen aus Daten zu optimieren. Diese Ansätze verbessern die Performance eines modernen Open-Source-MIP-Solvers von einer breiten Auswahl homogener und heterogener Instanzen. Im zweiten Teil dieser Arbeit leiten wir starke Schnittebenen für quadratischer Nebenbedingungen her. Durch Monoidal Strengthening verbessern wir Intersection Cuts, indem wir Ganzzahligkeit gezielt ausnutzen. Zudem zeigen wir, dass in unserem Setting Unique Lifting existiert, was impliziert, dass unser Verfahren zu den bestmöglichen Schnittkoeffizienten führt. Abschließend präsentieren wir ein allgemeines Framework zur Schnittebenenerzeugung, das Bedingungen identifiziert, unter denen eine Familie von Schnittebenen einen polyedrischen Abschluss ergibt. Damit können wir die Polyedrizität einer breiten Klasse gängiger Schnitte einfacher nachweisen."]},{"key":"dc:title","label":"Title","values":["Advancing mixed-integer programming using data-driven and deduction-based methods"]}]}],"canonical_facts":{"dc:contributor.advisor":["Pokutta, Sebastian"],"dc:creator":["Chmiela, Antonia"],"dc:date.accessioned":["2025-10-20T16:13:20Z"],"dc:date.available":["2025-10-20T16:13:20Z"],"dc:date.issued":["2025"],"dc:description.abstract":["Mixed-Integer Problems (MIPs) form one of the most general classes of optimization problems. As they are used to model many real-world scenarios, solving MIPs efficiently is crucial. Most solvers are based on the well-known Branch-and-Bound algorithm, which utilizes different subroutines to help find an optimal solution faster. In this dissertation, we focus on two of the most impactful components of MIP solving: Primal heuristics and cutting planes. First, we present two data-driven learning frameworks, offline and online, that aim to optimize the use of heuristics by learning from data describing their behavior. These approaches are able to improve performance of an state-of-the-art open-source MIP solver on a broad set of homogeneous and heterogeneous instances. In the second part of this thesis, we derive strong cutting planes to enforce quadratic constraints present in a MIP. By applying monoidal strengthening, we strengthen intersection cuts by exploiting integrality information. In addition, we show that, in our setting, unique lifting exists, implying that our strengthening procedure leads to the strongest cut coefficients. Finally, we present a general framework for cut generation to identify conditions under which a family of cutting planes yield a polyhedral closure. This allows us to show polyhedrality for a broad range of popular cuts more easily.","Gemischt-ganzzahlige Probleme (MIPs) bilden eine der allgemeinsten Klassen von Optimierungsproblemen. Da sie zur Modellierung vieler realer Szenarien verwendet werden, ist ihre effiziente Lösung von zentraler Bedeutung. Die meisten Lösungsverfahren basieren auf dem bekannten Branch-and-Bound-Algorithmus, der verschiedene Subroutinen nutzt, um schneller eine optimale Lösung zu finden. In dieser Dissertation konzentrieren wir uns auf zwei der wirkungsvollsten Komponenten: Primale Heuristiken und Schnittebenen. Zunächst stellen wir zwei datengetriebene Lernframeworks vor – eines im Offline- und eines im Online-Kontext – die darauf abzielen, den Einsatz von Heuristiken durch Lernen aus Daten zu optimieren. Diese Ansätze verbessern die Performance eines modernen Open-Source-MIP-Solvers von einer breiten Auswahl homogener und heterogener Instanzen. Im zweiten Teil dieser Arbeit leiten wir starke Schnittebenen für quadratischer Nebenbedingungen her. Durch Monoidal Strengthening verbessern wir Intersection Cuts, indem wir Ganzzahligkeit gezielt ausnutzen. Zudem zeigen wir, dass in unserem Setting Unique Lifting existiert, was impliziert, dass unser Verfahren zu den bestmöglichen Schnittkoeffizienten führt. Abschließend präsentieren wir ein allgemeines Framework zur Schnittebenenerzeugung, das Bedingungen identifiziert, unter denen eine Familie von Schnittebenen einen polyedrischen Abschluss ergibt. Damit können wir die Polyedrizität einer breiten Klasse gängiger Schnitte einfacher nachweisen."],"dc:identifier.uri":["https://depositonce.tu-berlin.de/handle/11303/25218","https://doi.org/10.14279/depositonce-24040"],"dc:language.iso":["en"],"dc:rights.uri":["https://creativecommons.org/licenses/by/4.0/"],"dc:title":["Advancing mixed-integer programming using data-driven and deduction-based methods"],"dc:type":["Doctoral Thesis"]},"updated_at":"2026-07-27T21:28:37Z"}