scientific article; zbMATH DE number 3910154

From MaRDI portal
Publication:3686448





Considered is the problem \[ (P)\quad \min \psi (x)\quad s.t.\quad x\in C,\quad h_ i(x)\leq 0\quad (i=1,...,m),\quad \psi (x)=[\frac{f_ 1(x)}{g_ 1(x)},...,\frac{f_ n(x)}{g_ n(x)}], \] where \((f_ 1,...,f_ n): X\to {\mathbb{R}}^ n\) is a continuous convex map, \((g_ 1,...,g_ n): X\to {\mathbb{R}}^ n\) a continuous concave map, \(h_ i: X\to {\mathbb{R}}\), \(i=1,...,m\), are continuous convex functions, not necessarily differentiable, X a real locally convex space, C a convex subspace of X. In Section 1 the results of \textit{P. Kanniappan} [J. Optimization Theory Appl. 40, 167-174 (1983; Zbl 0488.49007)] concerning necessary conditions of Fritz John and Kuhn-Tucker-type for Pareto optimality of nondifferentiable convex multiobjective programming are extended to the class of noncovex, nondifferentiable multiobjective programming problems (P). In Section 2 duality considerations with respect to (P) with corresponding results are presented. The dual to (P) is \[ (D)\quad \max \Phi (u,v,y,t)=V\quad subject\quad to \] \[ 0\in \partial F_ j(y,t)- v_ j\partial G_ j(y,t)+\sum^{m}_{i=1}u_ i\partial H_ i(y,t)+N_ T(y,t), \] \[ j=1,...,n,\quad t>0,\quad u_ i\geq 0,\quad i=1,...,m,\quad V\in {\mathbb{R}}^ n, \] where \(\partial f\) is the subdifferential of f (i.e. the set of all subgradients of f at \(x_ 0)\). A weak duality theorem is proved and the duality theorem is formulated and also proved: Theorem (duality): If \(x^ 0\) is any feasible solution of (P) and \((u^ 0,v^ 0,y^ 0,t^ 0)\) is any feasible solution of (D) such that \(\psi (x^ 0)=\Phi (u^ 0,v^ 0,y^ 0,t^ 0)\) then \(x^ 0\) is efficient for (P) and \((u^ 0,v^ 0,y^ 0,t^ 0)\) is efficient for (D).











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3686448)