U
Ufuk Topcu
Researcher at University of Texas at Austin
Publications - 504
Citations - 11791
Ufuk Topcu is an academic researcher from University of Texas at Austin. The author has contributed to research in topics: Markov decision process & Computer science. The author has an hindex of 44, co-authored 437 publications receiving 9636 citations. Previous affiliations of Ufuk Topcu include Google & University of Illinois at Urbana–Champaign.
Papers
More filters
Posted Content
Safe Reinforcement Learning via Shielding
TL;DR: A new approach to learn optimal policies while enforcing properties expressed in temporal logic by synthesizing a reactive system called a shield, which monitors the actions from the learner and corrects them only if the chosen action causes a violation of the specification.
Proceedings Article
Safe Reinforcement Learning via Shielding
TL;DR: In this paper, a reactive system called a shield is proposed to monitor the actions from the learner and correct them only if the chosen action causes a violation of the temporal logic specification, thus preserving the convergence guarantees of the learning system.
Journal ArticleDOI
Brief paper: Local stability analysis using simulations and sum-of-squares programming
TL;DR: A methodology utilizing information from simulations to generate Lyapunov function candidates satisfying necessary conditions for bilinear constraints is proposed and Qualified candidates are used to compute invariant subsets of the region-of-attraction and to initialize various bil inear search strategies for further optimization.
Proceedings ArticleDOI
Receding horizon control for temporal logic specifications
TL;DR: A receding horizon framework that satisfies a class of linear temporal logic specifications sufficient to describe a wide range of properties including safety, stability, progress, obligation, response and guarantee is described.
Journal ArticleDOI
Optimal power flow with large-scale storage integration
Dennice F. Gayme,Ufuk Topcu +1 more
TL;DR: In this paper, the authors investigate distributed energy storage as one such strategy, which augments the usual formulation by adding simple charge/discharge dynamics for energy storage collocated with load and/or generation buses cast as a finite-time optimal control problem.