English

Stochastic Three Points Method for Unconstrained Smooth Minimization

Optimization and Control 2019-05-08 v3

Abstract

In this paper we consider the unconstrained minimization problem of a smooth function in Rn{\mathbb{R}}^n in a setting where only function evaluations are possible. We design a novel randomized derivative-free algorithm --- the stochastic three points (STP) method --- and analyze its iteration complexity. At each iteration, STP generates a random search direction according to a certain fixed probability law. Our assumptions on this law are very mild: roughly speaking, all laws which do not concentrate all measure on any halfspace passing through the origin will work. For instance, we allow for the uniform distribution on the sphere and also distributions that concentrate all measure on a positive spanning set. Given a current iterate xx, STP compares the objective function at three points: xx, x+αsx+\alpha s and xαsx-\alpha s, where α>0\alpha>0 is a stepsize parameter and ss is the random search direction. The best of these three points is the next iterate. We analyze the method STP under several stepsize selection schemes (fixed, decreasing, estimated through finite differences, etc). We study non-convex, convex and strongly convex cases.

Keywords

Cite

@article{arxiv.1902.03591,
  title  = {Stochastic Three Points Method for Unconstrained Smooth Minimization},
  author = {El Houcine Bergou and Eduard Gorbunov and Peter Richtárik},
  journal= {arXiv preprint arXiv:1902.03591},
  year   = {2019}
}
R2 v1 2026-06-23T07:36:57.573Z