Computational Optimization and Applications
A stable primal---dual approach for linear programming under nondegeneracy assumptions
Computational Optimization and Applications
Improving the accuracy of linear programming solvers with iterative refinement
Proceedings of the 37th International Symposium on Symbolic and Algebraic Computation
| Hi-index | 0.00 |
The modern theory of condition measures for convex optimization problems was initially developed for convex problems in the conic format \[ (\mathrm{CP}_d)\qquad \ z_*:= \min_x \{c^t x | Ax-b \in C_Y, x \in C_X \}, \] and several aspects of the theory have now been extended to handle nonconic formats as well. In this theory, the (Renegar) condition measure $C(d)$ for $({\rm CP}_d)$ has been shown to be connected to bounds on a wide variety of behavioral and computational characteristics of $({\rm CP}_d)$, from sizes of optimal solutions to the complexity of algorithms for solving $({\rm CP}_d)$. Herein we test the practical relevance of the condition measure theory, as applied to linear optimization problems that one might typically encounter in practice. Using the NETLIB suite of linear optimization problems as a test bed, we found that 71% of the NETLIB suite problem instances have infinite condition measure. In order to examine condition measures of the problems that are the actual input to a modern interior-point-method (IPM) solver, we also computed condition measures for the NETLIB suite problems after preprocessing by CPLEX 7.1. Here we found that 19% of the postprocessed problem instances in the NETLIB suite have infinite condition measure, and that log C(d) of the postprocessed problems is fairly nicely distributed. Furthermore, among those problem instances with finite condition measure after preprocessing, there is a positive linear relationship between IPM iterations and log C(d) of the postprocessed problem instances (significant at the 95% confidence level), and 42% of the variation in IPM iterations among these NETLIB suite problem instances is accounted for by log C(d) of the postprocessed problem instances.