On the existence and the stability of solutions in nonconvex vector optimization †

Tran Van Nghi, Le Ngoc Kien, Nguyen Van Tuyen · 2025

Read On the existence and the stability of solutions in nonconvex vector optimization † by Tran Van Nghi, Le Ngoc Kien, Nguyen Van Tuyen (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-17974.txt.

Introduction

Tran Van Nghi and co-authors study in 2025 when solutions exist and stay stable in nonconvex vector optimization. Such problems weigh several aims at once, and their feasible sets can stretch without bound, which breaks tools built for tidy cases. The authors work with weak Pareto solutions, points where no aim can improve without hurting another.

Using cones and generalized functions that capture behavior far out, they give conditions for such solutions to exist and for a sharp minima property at infinity. They then show those conditions help judge stability when the problem is nudged by a linear change. A use case covers a subclass of robustly quasiconvex problems, where aims hold a looser form of convexity. The work gives analysts firmer ground for multi aim problems on open ended sets.

Catalog record

CREATORTran Van Nghi, Le Ngoc Kien, Nguyen Van Tuyen
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE52 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2502-17974

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

On the existence and the stability of solutions in nonconvex vector optimization † Tran Van Nghi, Le Ngoc Kien, Nguyen Van Tuyen Abstract The paper is devoted to the existence of weak Pareto solutions and the weak sharp minima at infinity property for a general class of constrained nonconvex vector optimization problems with unbounded constraint set via asymptotic cones and generalized asymptotic functions. Then we show that these conditions are useful for studying the solution stability of nonconvex vector optimization problems with linear perturbation. We also provide some applications for a subclass of robustly quasiconvex vector optimization problems. ∎ MSC 90C29 49J30 90C31 90C26 49J52 ——————————— Communicated by Hong-Kun Xu. † Dedicated to Professor Q.H. Ansari on the occasion of his 65th birthday 1 Introduction Vector optimization is an important part of mathematical programming. It is used to denote a type of optimization problems, where two or more objectives are to be minimized over a given set. Vector optimization problems arise naturally in a variety of theoretical and practical problems and have a wide range of applications. These problems are important tools for problems in economics, management, multicriteria design optimization, water resource planning, medicine, mathematical biology, etc; see, for example, [ 4 , 5 , 38 , 13 , 14 , 31 , 41 ] . An important and fundamental issue in theory of vector optimization problems is the investigation of existence and stability of solutions; see, for example, [ 3 , 14 , 38 , 31 , 11 , 30 , 39 , 23 , 24 , 25 ] . In this paper, we consider the following vector optimization problem \mathrm{Min}_{\mathbb{R}^{m}_{+}}\,\{f(x):=(f_{1}(x),\ldots,f_{m}(x))\;:\;x\in X\}, (VP) where X is a closed and unbounded subset in \mathbb{R}^{n} and f_{i}\colon\mathbb{R}^{n}\to\mathbb{R} , i~=~1,\ldots,m , are lower semicontinuous functions. Some authors have studied theorems of Frank–Wolfe type for polynomial vector optimization problems. Kim et al. [ 32 ] showed the existence of Pareto solutions of an unconstrained polynomial vector optimization problem under the assumptions that the Palais–Smale type condition holds and the image of the objective function has a bounded section. For the case where X is a convex semi-algebraic set and f is convex, Lee et al. [ 37 ] proved that ( VP ) has a Pareto solution if and only if f(X) has a nonempty bounded section. Recently, Flores-Bazán et al. [ 18 ] introduced two notions of vector asymptotic functions for vector functions with their image space ordered by the nonnegative orthant and established coercivity properties, coercive and noncoercive existence results for solutions of vector optimization problems. The main difficulty in investigation the existence of (weak) Pareto solutions is due to directions in unbounded constraint set along which the objective function may decrease. Asymptotic tools have proved to be very useful in order to overcome the above difficulty; see, for example, [ 45 , 48 , 1 , 18 , 36 , 9 ] . The idea of the asymptotic approach is based on the asymptotic behavior of sets and functions. When the constraint set is unbounded, asymptotic tools allow us to obtain coercivity properties and existence results for vector optimization problems. In this paper, we investigate the existence and the stability of weak Pareto solutions of constrained nonconvex vector optimization problems by using the tool of asymptotic analysis. We make the following contributions to vector optimization problems: \bullet As a first result, we present new results on existence of weak Pareto solutions and the weak sharp minima at infinity property for a general class of constrained nonconvex vector optimization problems with unbounded constraint set via asymptotic cones and generalized asymptotic functions. \bullet We obtain stability results, including: the nonemptiness and compactness, the upper/lower semicontinuity, and the closedness of the weak Pareto solution map of the considered problem under linear perturbations. \bullet We introduce the concept of robustly quasiconvex for vector-valued functions and apply our previous results to the particular case when the objective function is quasiconvex or \alpha -robustly quasiconvex. By using the q -asymptotic function, we develop sufficient conditions for the mentioned problems under weaker assumptions. This paper is organized as follows. Section 2 provides the necessary background required for the subsequent development. In Section 3 , we investigate the solution existence and the weak sharp minima at infinity. Section 4 presents stability results under a linear perturbation. Section 5 proposes some applications to the particular case when the objective function is quasiconvex or \alpha -robustly quasiconvex. A summary of conclusions is explained in the last section. 2 Preliminaries Throughout the paper, the considered spaces are finite-dimensional Euclidean with the inner product and the norm being denoted, respectively, by \langle\cdot,\cdot\rangle and \|\cdot\| . The open ball centered at the origin with a radius of \alpha>0 and the set of all positive integer numbers are denoted, respectively, by \mathbb{B}_{\alpha} and \mathbb{N} . The nonnegative orthant of \mathbb{R}^{m} is defined by \mathbb{R}^{m}_{+}:=\{x=(x_{1},\ldots,x_{m})\in\mathbb{R}^{m}\;:\;x_{i}\geq 0,i=1,\ldots,m\}. Let \varphi\colon\mathbb{R}^{n}\rightarrow\overline{\mathbb{R}}:=\mathbb{R}\cup\{\pm\infty\} be an extended-real-valued function. The domain and epigraph of \varphi are defined, respectively, by \displaystyle\operatorname{dom}\varphi:=\{x\in\mathbb{R}^{n}\,:\,\varphi(x)<+\infty\}\ \ \text{and} \displaystyle\mathrm{epi}\,\varphi:=\{(x,t)\in\mathrm{dom}\,\varphi\times\mathbb{R}\,:\,\varphi(x)\leq t\}. We say that \varphi is proper if \operatorname{dom}\varphi\neq\emptyset and \varphi(x)>-\infty for every x\in\mathbb{R}^{n} . We adopt the usual convention \inf\nolimits_{\emptyset}\varphi:=+\infty and \sup\nolimits_{\emptyset}\varphi:=-\infty . For a given \alpha\in\mathbb{R} , we denote by \mathrm{lev}(\varphi,\alpha):=\{x\in\mathbb{R}^{n}:~\varphi(x)\leq\alpha\} the sublevel set of \varphi at value \alpha . For a nonempty subset X in \mathbb{R}^{n} , the set of all minimum points of \varphi on X is defined by \mathrm{argmin}_{X}\varphi:=\{x\in X:~\varphi(x)\leq\varphi(y)\ \ \forall y\in X\}. Definition 1 We say that the function \varphi is: (i) convex if \mathrm{epi}\,\varphi is a convex subset of \mathbb{R}^{n}\times\mathbb{R} . (ii) quasiconvex if \mathrm{lev}(\varphi,\alpha) is a convex subset of \mathbb{R}^{n} for all \alpha\in\mathbb{R} . It is well-known that the function \varphi is convex if and only if its domain is convex and for every x,y\in{\rm dom}\,\varphi \varphi(\lambda x+(1-\lambda)y)\leq\lambda\varphi(x)+(1-\lambda)\varphi(y)\ \ \forall\lambda\in[0,1] and it is quasiconvex if and only if for every x,y\in{\rm dom}\,\varphi , \varphi(\lambda x+(1-\lambda)y)\leq\max\{\varphi(x),\varphi(y)\}\ \ \forall\lambda\in[0,1]. Clearly, every convex function is also quasiconvex, but not vice versa. For example, the function \varphi(x):=\sqrt{|x|} for all x\in\mathbb{R} is quasiconvex but it is nonconvex. In ( 43 , Lemma 1.1) , the authors showed that the quasiconvexity is not stable under linear perturbations and hence, they introduced the so-called \alpha -robustly quasiconvex functions. Definition 2 ( see [ 10 , 43 ] ) For \alpha\geq 0 , a proper function \varphi:\mathbb{R}^{n}\rightarrow\overline{\mathbb{R}} is said to be \alpha -robustly quasiconvex if the function x\mapsto\varphi(x)+\langle u,x\rangle is quasiconvex for all u\in\mathbb{B}_{\alpha} . As shown in ( 10 , p. 1091) , every convex function is also \alpha -robustly quasiconvex for all \alpha\geq 0 , but not vice versa. Furthermore, every local minimum of an \alpha -robustly quasiconvex function is a global one. For a further study we refer the readers to [ 10 , 43 , 44 ] . We now introduce the concept of robustly quasiconvex for vector-valued functions. Definition 3 Let f\colon\mathbb{R}^{n}\to\mathbb{R}^{m} , x\mapsto f(x):=(f_{1}(x),\ldots,f_{m}(x)) be a mapping and \alpha\geq 0 . We say that f is: (i) \mathbb{R}^{m}_{+} -quasiconvex if for every y=(y_{1},\ldots,y_{m})\in\mathbb{R}^{m} , the sublevel set \mathrm{lev}(f,y):=\{x\in\mathbb{R}^{n}\,:\,f_{i}(x)\leq y_{i},\ \ i=1,\ldots,m\} of f at y is a convex subset of \mathbb{R}^{n} . (ii) \mathbb{R}^{m}_{+} - \alpha -robustly quasiconvex if the mapping x\mapsto f(x)+\langle u,x\rangle is \mathbb{R}^{m}_{+} - quasiconvex for all u=(u_{1},\ldots,u_{m})\in\mathbb{R}^{m\times n}

Read the complete text