First- and second-order optimality conditions for multiobjective fractional programming (Q2355009)

The authors consider the nonsmooth multiobjective fractional programming problem consisting in minimizing \(\varphi \left( x\right) :=(\frac{f_{1}(x)}{g_{1}(x) },\dots,\frac{f_{m}(x)}{g_{m}(x)})\) subject to \(h(x)\in -K\), with real-valued functions \(f_{i}\) and \(g_{i}\) being defined on a normed space \(X\), \(h\) being a mapping from \(X\) into another normed space \(Y\), and \(K\) being a proper closed convex cone in \(Y\) with nonempty interior. The space \(\mathbb{R }^{m}\) is also endowed with the order induced by another such cone \(C\), assumed also to be pointed. The denominator functions \(g_{i}\) are assumed to be continuous and nonvanishing. The first- and second-order necessary conditions and sufficient conditions for optimality presented in this paper do not require any convexity assumption and are in terms of the notions of first- and second-order approximations introduced by \textit{A. Jourani} and \textit{L. Thibault} [Math. Oper. Res. 18, No. 2, 390--401 (1993; Zbl 0779.49021)], for which new properties and calculus rules are obtained. Illustrative examples are also provided.