243. Clarabel: An interior point solver for conic optimization
Invited abstract in session TC-2: Developments in interior point methods, stream Developments in interior point methods.
Thursday, 11:25 - 12:40Room: M:O
Authors (first author is the speaker)
| 1. | Yuwen Chen
|
| EPFL STI IGM LA3, EPFL |
Abstract
The homogeneous self-dual embedding (HSDE) is a popular technique that unifies optimality and infeasibility detection of a convex problem and has been widely used in convex interior-point solver. However, it only allows a linear cost in the objective and we need to eliminate the quadratic term in the objective function, replacing it with an epigraphical upper bound and an additional second-order cone constraint in the objective.
We presented the interior-point solver, Clarabel, for conic optimization with quadratic objectives by using the homogeneous embedding for infeasibility detection, which tackled the issue of transforming QPs into SOCPs. The solver is fully written in both Julia and Rust languages. We also support second-order cones, exponential cones, power cones and positive definite cones in Clarabel. Our solver is competitive with the cutting-edge conic solvers on various benchmarks and outperforms them in terms of time and numerical stability on QPs and SOCPs.
Keywords
- Conic and semidefinite optimization
- Analysis and engineering of optimization algorithms
Status: accepted
Back to the list of papers