Print Email Facebook Twitter From Supervised to Reinforcement Learning: an Inverse Optimization Approach Title From Supervised to Reinforcement Learning: an Inverse Optimization Approach Author Dimanidis, Ioannis (TU Delft Mechanical, Maritime and Materials Engineering) Contributor Mohajerin Esfahani, P. (mentor) Mazo, M. (graduation committee) Atasoy, B. (graduation committee) Degree granting institution Delft University of Technology Programme Mechanical Engineering | Systems and Control Date 2021-12-10 Abstract We propose a novel method combining elements of supervised- and Q-learning for the control of dynamical systems subject to unknown disturbances. By using the Inverse Optimization framework and in-hindsight information we can derive a causal parametric optimization policy that approximates a non-causal MPC expert. Furthermore, we propose a new min-max MPC scheme that robustifies against a ball around a disturbance trajectory. This scheme yields an exact convex reformulation using the S-Lemma, and is also approximated using Inverse Optimization. Finally, simulation studies clarify and verify our approach. Subject Inverse OptimizationReinforcement LearningConvex OptimizationOptimal Controldata-driven control To reference this document use: http://resolver.tudelft.nl/uuid:9d5efafe-58e5-497a-b7ea-0ac0ce2e9173 Part of collection Student theses Document type master thesis Rights © 2021 Ioannis Dimanidis Files PDF main.pdf 1.35 MB Close viewer /islandora/object/uuid:9d5efafe-58e5-497a-b7ea-0ac0ce2e9173/datastream/OBJ/view