Gradient-free methods for nonsmooth nonconvex multiobjective optimization
Optimization Eruditorum, Volume 2, Issue 3, December 2025, Pages 157–164
YONG ZHAO
College of Mathematics and Statistics, Chongqing Jiaotong University, Chongqing 400074, China
CHUN-LIN YOU
College of Mathematics and Statistics, Chongqing Jiaotong University, Chongqing 400074, China
WANG CHEN
National Center for Applied Mathematics in Chongqing, Chongqing Normal University, Chongqing 401331, China
Abstract
This paper introduces a gradient-free method for solving a class of nonsmooth nonconvex multiobjective optimization problems. Under standard regularity conditions, we establish a non-asymptotic convergence rate of the proposed algorithm, where optimality is quantified in terms of \((\delta, \epsilon)\)-Goldstein stationarity measure. Numerical experiments are presented to demonstrate the effectiveness of the proposed method.
Cite this Article as
Yong Zhao, Chun-Lin You, and Wang Chen, Gradient-free methods for nonsmooth nonconvex multiobjective optimization, Optimization Eruditorum, 2(3), 157–164, 2025