Nonlinear Derivative-free Constrained Optimization with a Mixed Penalty-Logarithmic Barrier Approach and Direct Search
Published in DFOS 2nd Derivative-free Optimization Symposium, 2024
DFOS 2nd Derivative-free Optimization Symposium
Department of Mathematics “Tullio Levi-Civita” University of Padova More information here
This is a single-track symposium, highlighting the latest trends in Derivative-free Optimization (DFO), with invited speakers only. The main goal is to find new connections between research focused in DFO algorithmic design and novel applications of DFO. A single-track workshop facilitates the interaction among participants, strengthening possibilities of collaboration.
Nonlinear Derivative-free Constrained Optimization with a Mixed Penalty-Logarithmic Barrier Approach and Direct Search
Abstract: We propose the use of a mixed penalty-logarithmic barrier approach and direct search for addressing nonlinearly constrained DFO problems. Inequality constraints are splitted in two sets: one treated with the logarithmic barrier approach, another addressed with a penalization term, along with the equality constraints. Under smooth assumptions, convergence is established, and numerical experiments show the robustness and efficiency of the proposed method, when compared to state-of-the-art solvers.
Joint work with Andrea Brilli, Ana Luísa Custódio and Giampaolo Liuzzi.
Acknowledgments: This research was financially supported by Fundação para a Ciência e a Tecnologia (FCT) (Portuguese Foundation for Science and Technology) through projects UIDB/00297/2020, UIDP/00297/2020, and UI/BD/151246/2021 (Centro de Matemática e Aplicações - NOVA Math).