Mixed duality without a constraint qualification for minimax fractional programming

From MaRDI portal
Publication:4467183





Considered is the following minimax fractional programming problem: \[ \text{(P)}\qquad \qquad \min_{x\in S}~\max_{1\leq i\leq p}[f_{i}(x)/g_{i}(x)] \] where \(S=\left\{ x\in \mathbb R^{n}~| ~h_{j}(x)\leq 0,~j=1,2,...,m\right\} ,~f_{i}-g_{i},~h_{j}:\mathbb R^{n}\rightarrow \mathbb R\) are differentiable pseudoconvex functions and \(f_{i}(x)\geq 0,~g_{i}(x)>0\). Without any constraint qualification, the authors establish the necessary and sufficient optimality conditions for Problem (P). They use these optimality conditions to formulate a mixed dual model which unifies the Mond-Weir dual, Wolfe dual and parameter dual models. Weak, strong and strict converse dual theorems are proved.











This page was built for publication: Mixed duality without a constraint qualification for minimax fractional programming

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4467183)