arXiv: Scaly: an Optimizing Compiler for Optimal Control

arXiv: Scaly: an Optimizing Compiler for Optimal Control

Tudor A. Oancea and Colin N. Jones present Scaly, an open-source optimizing compiler written in Python that generates C code for optimal control. It provides a modeling language embedded in Python for parametric optimization problems used in model predictive control and moving horizon estimation, and computes sparse derivatives that standard optimization solvers need. The authors’ central design choice is to preserve repeated structure in an optimal-control problem, such as the stages of a horizon, through differentiation into generated loops. They say the generated C depends only on the C math library and never allocates memory dynamically. On three benchmark problems, Scaly evaluates the Hessian of the Lagrangian up to 26 times faster than CasADi, with smaller binaries and build times that stay nearly constant with the horizon. With the same solver, the authors report shorter solver time per control step on every benchmark, by up to a factor of 9. These results describe performance on the tested problems; the supplied abstract does not give benchmark details or the experimental setup.

Publication: 4380da54-637e-4c6f-82eb-dae1b7a893f0

attribution — pass: Authors Tudor A. Oancea and Colin N. Jones are correctly named, and arXiv is correctly identified as the venue. Most performance claims are framed as author reports. One small repair: the summary sentence on the 26x Hessian speedup over CasADi is stated in Scaly's own voice. Prefixing it with 'the authors report' would keep it consistent with the claims list.

Authors:Tudor A. Oancea, Colin N. Jones

summary — uncertain: The summary is faithful but hard to follow without specialist background. Central terms go unexplained or lack tooltips: optimal control, model predictive control, moving horizon estimation, Hessian of the Lagrangian, sparse derivatives. CasADi is never identified as the comparison tool. The summary also does not say plainly why faster solving per control step matters, for example that controllers must compute decisions repeatedly in real time. Smallest repair: add term-library tooltips for these terms, identify CasADi as an existing tool, and add one sentence on the significance the source supports.

computes the sparse derivatives that standard optimization solvers need

date — pass: The displayed original publication date of 2026-10-30 matches the v1 submission date on arXiv. The later feed date of 2026-11-02 is correctly not used.

[Submitted on 30 Oct 2026]

support — pass: Every claim matches the abstract verbatim. Calling structure preservation the authors' 'central design choice' is a mild framing, but the abstract supports it. The limitations and the question correctly scope the results to three benchmarks and add no unsupported benefits or costs.

With the same solver, Scaly shortens the solver time per control step on every benchmark, by a factor of up to 9.

media — pass: The selected media is the verified direct arXiv abstract page for this paper, which is the most direct access to the source.