On sufficiency and duality for generalized quasiconvex programs (Q1323225): Difference between revisions

Sufficient optimality conditions and duality theorems for nonlinear optimization problems with inequality and equality constraints are presented. The classical notions of convexity, quasiconvexity and pseudoconvexity are replaced by the more general notions of \(({\mathcal F},\rho)\)-convexity, \(({\mathcal F},\rho)\)-quasiconvexity and \(({\mathcal F},\rho)\)-pseudoconvexity respectively. So for example a differentiable function \(f: \mathbb{R}^ n\to \mathbb{R}\) is called \(({\mathcal F},\rho)\)-convex if \[ f(x)- f(x^ 0)\geq F(x,x^ 0;\nabla f(x^ 0))+ \rho d^ 2(x,x^ 0), \] where \(F(x,x^ 0;.): \mathbb{R}^ n\to \mathbb{R}\) is a sublinear function, \(\rho\in \mathbb{R}\) and \(d(.,.): \mathbb{R}^ n\times \mathbb{R}^ n\to \mathbb{R}\) a function with \(d(x,x)= 0\) for any \(x\in \mathbb{R}^ n\). Naturally, if \(F(x,x^ 0;.)\) is linear and \(\rho= 0\) then \(f\) is convex. In the same manner the other notions are introduced. With this generalizations the author is able to construct a sufficient condition for \(x^ 0\) to be a global solution of the given optimization problem. Moreover, for the appropriate Mond-Weir dual problem a direct and an inverse duality assertion is presented.