{"id":{"repo_id":"iastate","oai_identifier":"oai:dr.lib.iastate.edu:20.500.12876/106702"},"canonical_url":"https://search.dev.ndltd.org/etd/iastate/oai:dr.lib.iastate.edu:20.500.12876/106702","repository":{"repo_id":"iastate","name":"Iowa State University","base_url":"https://dr.lib.iastate.edu/server/oai/request"},"display":{"title":"Machine learning-aided trajectory optimization and tour design","abstract":"This dissertation investigates machine learning (ML)-aided methods for preliminary space mission design, with a focus on trajectory optimization and tour design for interplanetary missions involving multiple gravity assists. In early-stage trajectory analysis, the repeated solution of Lambert’s problem often becomes a major computational bottleneck. Meanwhile, the number of possible flyby sequences grows combinatorially as the number of flybys increases, and each sequence is associated with a distinct cost function topology, making optimization and trade-space exploration especially challenging. To address these issues, this dissertation develops and evaluates several ML-aided frameworks for rapid transfer approximation, flyby-sequence optimization, and reinforcement-learning-based trajectory optimization. First, a neural-network-based Lambert approximator is developed to provide rapid estimates of transfer legs needed for large-scale trajectory searches while maintaining sufficient accuracy for preliminary mission design. Second, the neural network model for Lambert’s problem is integrated into the dynamic-size multiple-populations genetic algorithm (DSMPGA), and a vectorized implementation is adopted to enhance computational speed. This approach handles variable-length gravity-assist trajectories and generates both optimal and suboptimal mission candidates. Third, reinforcement learning is investigated as a sequential decision-making framework for multiple-gravity-assist trajectory optimization. In addition, deep learning is investigated for the placement of deep-space maneuvers, and the Extreme Theory of Functional Connections (X-TFC) is applied to orbital transfer problems under non-Keplerian dynamics, including J2-perturbed two-body motion and the circular restricted three-body problem. Overall, the results demonstrate that ML can accelerate preliminary trajectory design, improve the exploration of large and complex search spaces, and provide a promising foundation for future ML-aided mission design frameworks in space exploration.","abstract_html":"This dissertation investigates machine learning (ML)-aided methods for preliminary space mission design, with a focus on trajectory optimization and tour design for interplanetary missions involving multiple gravity assists. In early-stage trajectory analysis, the repeated solution of Lambert’s problem often becomes a major computational bottleneck. Meanwhile, the number of possible flyby sequences grows combinatorially as the number of flybys increases, and each sequence is associated with a distinct cost function topology, making optimization and trade-space exploration especially challenging. To address these issues, this dissertation develops and evaluates several ML-aided frameworks for rapid transfer approximation, flyby-sequence optimization, and reinforcement-learning-based trajectory optimization. First, a neural-network-based Lambert approximator is developed to provide rapid estimates of transfer legs needed for large-scale trajectory searches while maintaining sufficient accuracy for preliminary mission design. Second, the neural network model for Lambert’s problem is integrated into the dynamic-size multiple-populations genetic algorithm (DSMPGA), and a vectorized implementation is adopted to enhance computational speed. This approach handles variable-length gravity-assist trajectories and generates both optimal and suboptimal mission candidates. Third, reinforcement learning is investigated as a sequential decision-making framework for multiple-gravity-assist trajectory optimization. In addition, deep learning is investigated for the placement of deep-space maneuvers, and the Extreme Theory of Functional Connections (X-TFC) is applied to orbital transfer problems under non-Keplerian dynamics, including J2-perturbed two-body motion and the circular restricted three-body problem. Overall, the results demonstrate that ML can accelerate preliminary trajectory design, improve the exploration of large and complex search spaces, and provide a promising foundation for future ML-aided mission design frameworks in space exploration.","abstract_has_math":false,"creators":["Choi, Sungmoon"],"institution":"Iowa State University - Thesis & Dissertation","degree_name":"Doctor of Philosophy","degree_level":"dissertation","degree_discipline":"Aerospace engineering","degree_department":null,"school":null,"contributors":[],"advisors":["Abdelkhalik, Ossama"],"committee_chairs":[],"committee_members":[],"year":2026,"date_issued":"2026-05","date_published":"2026-05","updated_at":"2026-07-24T02:37:27Z","subjects":["Aerospace engineering"],"languages":["en_US"],"rights":[],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"https://dr.lib.iastate.edu/handle/20.500.12876/106702","outbound_label":"Repository record","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Abdelkhalik, Ossama"]},{"key":"dc:creator","label":"Author","values":["Choi, Sungmoon"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2026-06-09T22:24:29Z"]},{"key":"dc:date.available","label":"Dc Date Available","values":["2026-06-09T22:24:29Z"]},{"key":"dc:date.issued","label":"Date","values":["2026-05"]},{"key":"dc:type","label":"Dc Type","values":["Text"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Aerospace engineering"]},{"key":"thesis:degree_level","label":"Degree Level","values":["dissertation"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Doctor of Philosophy"]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["Iowa State University - Thesis & Dissertation"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Aerospace engineering"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language.iso","label":"Language (ISO)","values":["en_US"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://dr.lib.iastate.edu/handle/20.500.12876/106702"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["May2026"]},{"key":"dc:description.abstract","label":"Abstract","values":["This dissertation investigates machine learning (ML)-aided methods for preliminary space mission design, with a focus on trajectory optimization and tour design for interplanetary missions involving multiple gravity assists. In early-stage trajectory analysis, the repeated solution of Lambert’s problem often becomes a major computational bottleneck. Meanwhile, the number of possible flyby sequences grows combinatorially as the number of flybys increases, and each sequence is associated with a distinct cost function topology, making optimization and trade-space exploration especially challenging. To address these issues, this dissertation develops and evaluates several ML-aided frameworks for rapid transfer approximation, flyby-sequence optimization, and reinforcement-learning-based trajectory optimization. First, a neural-network-based Lambert approximator is developed to provide rapid estimates of transfer legs needed for large-scale trajectory searches while maintaining sufficient accuracy for preliminary mission design. Second, the neural network model for Lambert’s problem is integrated into the dynamic-size multiple-populations genetic algorithm (DSMPGA), and a vectorized implementation is adopted to enhance computational speed. This approach handles variable-length gravity-assist trajectories and generates both optimal and suboptimal mission candidates. Third, reinforcement learning is investigated as a sequential decision-making framework for multiple-gravity-assist trajectory optimization. In addition, deep learning is investigated for the placement of deep-space maneuvers, and the Extreme Theory of Functional Connections (X-TFC) is applied to orbital transfer problems under non-Keplerian dynamics, including J2-perturbed two-body motion and the circular restricted three-body problem. Overall, the results demonstrate that ML can accelerate preliminary trajectory design, improve the exploration of large and complex search spaces, and provide a promising foundation for future ML-aided mission design frameworks in space exploration."]},{"key":"dc:format.mimetype","label":"Dc Format Mimetype","values":["PDF"]},{"key":"dc:title","label":"Title","values":["Machine learning-aided trajectory optimization and tour design"]}]}],"canonical_facts":{"dc:contributor.advisor":["Abdelkhalik, Ossama"],"dc:creator":["Choi, Sungmoon"],"dc:date.accessioned":["2026-06-09T22:24:29Z"],"dc:date.available":["2026-06-09T22:24:29Z"],"dc:date.issued":["2026-05"],"dc:description":["May2026"],"dc:description.abstract":["This dissertation investigates machine learning (ML)-aided methods for preliminary space mission design, with a focus on trajectory optimization and tour design for interplanetary missions involving multiple gravity assists. In early-stage trajectory analysis, the repeated solution of Lambert’s problem often becomes a major computational bottleneck. Meanwhile, the number of possible flyby sequences grows combinatorially as the number of flybys increases, and each sequence is associated with a distinct cost function topology, making optimization and trade-space exploration especially challenging. To address these issues, this dissertation develops and evaluates several ML-aided frameworks for rapid transfer approximation, flyby-sequence optimization, and reinforcement-learning-based trajectory optimization. First, a neural-network-based Lambert approximator is developed to provide rapid estimates of transfer legs needed for large-scale trajectory searches while maintaining sufficient accuracy for preliminary mission design. Second, the neural network model for Lambert’s problem is integrated into the dynamic-size multiple-populations genetic algorithm (DSMPGA), and a vectorized implementation is adopted to enhance computational speed. This approach handles variable-length gravity-assist trajectories and generates both optimal and suboptimal mission candidates. Third, reinforcement learning is investigated as a sequential decision-making framework for multiple-gravity-assist trajectory optimization. In addition, deep learning is investigated for the placement of deep-space maneuvers, and the Extreme Theory of Functional Connections (X-TFC) is applied to orbital transfer problems under non-Keplerian dynamics, including J2-perturbed two-body motion and the circular restricted three-body problem. Overall, the results demonstrate that ML can accelerate preliminary trajectory design, improve the exploration of large and complex search spaces, and provide a promising foundation for future ML-aided mission design frameworks in space exploration."],"dc:format.mimetype":["PDF"],"dc:identifier.uri":["https://dr.lib.iastate.edu/handle/20.500.12876/106702"],"dc:language.iso":["en_US"],"dc:subject":["Aerospace engineering"],"dc:title":["Machine learning-aided trajectory optimization and tour design"],"dc:type":["Text"],"thesis:degree_discipline":["Aerospace engineering"],"thesis:degree_level":["dissertation"],"thesis:degree_name":["Doctor of Philosophy"],"thesis:institution_name":["Iowa State University - Thesis & Dissertation"]},"updated_at":"2026-07-24T02:37:27Z"}