Performance enhancements for a generic conic interior point algorithm
Name
12532_2022_Article_226.pdf
Size
971.66 KB
Format
Adobe PDF
Checksum (MD5)
12df477599a09ec176172f89eb5262a6
Author(s)
Coey, Chris
Kapelevich, Lea
Vielma, Juan P.
Date Issued
September 17, 2022
Publisher
Springer Berlin Heidelberg
Citation
Coey, Chris, Kapelevich, Lea and Vielma, Juan P. 2022. "Performance enhancements for a generic conic interior point algorithm."
Version
Final published version
Abstract
Abstract
In recent work, we provide computational arguments for expanding the class of proper cones recognized by conic optimization solvers, to permit simpler, smaller, more natural conic formulations. We define an exotic cone as a proper cone for which we can implement a small set of tractable (i.e. fast, numerically stable, analytic) oracles for a logarithmically homogeneous self-concordant barrier for the cone or for its dual cone. Our extensible, open-source conic interior point solver, Hypatia, allows modeling and solving any conic problem over a Cartesian product of exotic cones. In this paper, we introduce Hypatia’s interior point algorithm, which generalizes that of Skajaa and Ye (Math. Program. 150(2):391–422, 2015) by handling exotic cones without tractable primal oracles. To improve iteration count and solve time in practice, we propose four enhancements to the interior point stepping procedure of Skajaa and Ye: (1) loosening the central path proximity conditions, (2) adjusting the directions using a third order directional derivative barrier oracle, (3) performing a backtracking search on a curve, and (4) combining the prediction and centering directions. We implement 23 useful exotic cones in Hypatia. We summarize the complexity of computing oracles for these cones and show that our new third order oracle is not a bottleneck. From 37 applied examples, we generate a diverse benchmark set of 379 problems. Our computational testing shows that each stepping enhancement improves Hypatia’s iteration count and solve time. Altogether, the enhancements reduce the geometric means of iteration count and solve time by over 80% and 70% respectively.
MIT Department
Massachusetts Institute of Technology. Operations Research Center
Sloan School of Management
Terms of Use
Creative Commons Attribution
Persistent DSpace Link
DOI of Published Version
https://doi.org/10.1007/s12532-022-00226-0
