A Sequential Quadratic Programming Method With High-Probability Complexity Bounds for Nonlinear Equality-Constrained Stochastic Optimization
Year: 2025
Type: article
Abstract: A step-search sequential quadratic programming method is proposed for solving nonlinear equality-constrained stochastic optimization problems. It is assumed that constraint function values and derivatives are available, but only stochastic approximations of the objective function and its associated derivatives can be computed via inexact probabilistic zeroth- and first-order oracles. Under reasonable assumptions, a high probability bound on the number of iterations that the algorithm requires to reach a first-order [Formula: see text]-stationary iterate is derived, where [Formula: see text] is lower bounded by a positive quantity dictated by the noise level of the inexact probabilistic zeroth- and first-order oracles. Numerical results on standard nonlinear optimization test problems illustrate the advantages and limitations of our proposed method.
Source: SIAM Journal on Optimization
DOI: 10.1137/23m1549006
Authors: Albert S. Berahas, Miaolan Xie, Baoyu Zhou
Language: English
Subfield: Numerical Analysis
Field: Mathematics
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W4406334657
