Inexact proximal point method with a Bregman regularization for quasiconvex multiobjective optimization problems via limiting subdifferentials

Abstract

In this paper, we investigate a class of unconstrained multiobjective optimization problems (abbreviated as, MPQs), where the components of the objective function are locally Lipschitz and quasiconvex. To solve MPQs, we introduce an inexact proximal point method with Bregman distances (abbreviated as, IPPMB) via Mordukhovich limiting subdifferentials. We establish the well-definedness of the sequence generated by the IPPMB algorithm. Based on two versions of error criteria, we introduce two variants of IPPMB, namely, IPPMB1 and IPPMB2. Moreover, we establish that the sequences generated by the IPPMB1 and IPPMB2 algorithms converge to the Pareto–Mordukhovich critical point of the problem MPQ. In addition, we derive that if the components of the objective function of MPQ are convex, then the sequences converge to the weak Pareto efficient solution of MPQ. Furthermore, we discuss the linear and superlinear convergence of the sequence generated by the IPPMB2 algorithm. We furnish several non-trivial numerical examples to demonstrate the effectiveness of the proposed algorithms and solve them by employing MATLAB R2023b. To demonstrate the applicability and significance of the IPPMB algorithm, we solve a nonsmooth large-scale sparse quasiconvex multiobjective optimization by employing MATLAB R2023b.