Deep FlexQP: Accelerated Nonlinear Programming via Deep Unfolding
Alex Oshin [email protected] Institute of TechnologyRahul Vodeb Ghosh [email protected] Institute of Technology, Augustinos D. Saravanos [email protected] Institute of TechnologyEvangelos A. Theodorou [email protected] Institute of Technology · 2025
Read Deep FlexQP: Accelerated Nonlinear Programming via Deep Unfolding by Alex Oshin [email protected] Institute of TechnologyRahul Vodeb Ghosh [email protected] Institute of Technology, Augustinos D. Saravanos [email protected] Institute of TechnologyEvangelos A. Theodorou [email protected] Institute of Technology (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2512-01565.txt.
Description
Alex Oshin and co-authors speed up nonlinear programming with deep unfolding. Quadratic programs sit inside many control and planning tasks. Solvers must stay feasible at each step, which can be slow. Their solver, FlexQP, is built to stay feasible by design.
Deep unfolding turns solver steps into a learnable network. Training tunes those steps for speed on real task families. An always-feasible solver can run in tight loops, like robot control. Learned speed with firm feasibility could cut compute in autonomous systems. Unfolded steps keep solver structure, so the learned solver stays readable to control engineers who must certify it before real use.
Catalog record
| CREATOR | Alex Oshin [email protected] Institute of TechnologyRahul Vodeb Ghosh [email protected] Institute of Technology, Augustinos D. Saravanos [email protected] Institute of TechnologyEvangelos A. Theodorou [email protected] Institute of Technology |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 94 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2512-01565