The Minimizer of the Sum of Two Strongly Convex Functions
From MaRDI portal
Publication:6437508
arXiv2305.13134MaRDI QIDQ6437508FDOQ6437508
Authors: Kananart Kuwaranancharoen, Shreyas Sundaram 
Publication date: 22 May 2023
Abstract: The problem of finding the minimizer of a sum of convex functions is central to the field of optimization. In cases where the functions themselves are not fully known (other than their individual minimizers and convexity parameters), it is of interest to understand the region containing the potential minimizers of the sum based only on those known quantities. Characterizing this region in the case of multivariate strongly convex functions is far more complicated than the univariate case. In this paper, we provide both outer and inner approximations for the region containing the minimizer of the sum of two strongly convex functions, subject to a constraint on the norm of the gradient at the minimizer of the sum. In particular, we explicitly characterize the boundary and interior of both outer and inner approximations. Interestingly, the boundaries as well as the interiors turn out to be identical and we show that the boundary of the region containing the potential minimizers is also identical to that of the outer and inner approximations.
Convex programming (90C25) Quadratic and bilinear forms, inner products (15A63) Convexity of real functions of several variables, generalizations (26B25) Convex functions and convex programs in convex geometry (52A41)
This page was built for publication: The Minimizer of the Sum of Two Strongly Convex Functions
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6437508)
