Nondifferentiable fractional programming in complex spaces involving \(({\mathfrak L},\varrho,\theta)\)-convex analytic functions (Q698598)

The authors consider the following nondifferentiable fractional programming problem in the complex space. \[ \min{\text{Re}[f(z,\overline z)+ (z^H Az)^{1/2}]\over \text{Re}[g(z,\overline z)- (z^H Bz)^{1/2}]}= {\text{Re }F(z,\overline z)\over \text{Re }G(z,\overline z)}\tag{P} \] s.t. \(z\in\mathbb{C}^n\) and \(h(z,\overline z)\in S\subset \mathbb{C}^m\), where \(f,g: \mathbb{C}^n\times \mathbb{C}^n\to\mathbb{C}\) and \(h: \mathbb{C}^n\times \mathbb{C}^n\to \mathbb{C}^m\) are analytic functions, \(A,B\in \mathbb{C}^{n\times n}\) are positive semidefinite Hermitian matrices and \(S\) is a polyhedral cone in \(\mathbb{C}^m\). It is assumed that \(\text{Re }F(z,\overline z)\geq 0\) and \(\text{Re }G(z,\overline z)> 0\) for all feasible solution \((z,\overline z)\) to problem (P). The authors establish optimality conditions for the problem (P) involving generalized \(({\mathcal L},\rho,\theta)\)-convexity for the functions \(F\), \(G\) and \(h\). They construct a dual model to problem (P) and establish the weak, strong and strict converse duality theorems. The results represent a generalization of the results obtained by \textit{O. Parkash}, \textit{P. C. Saxena} and \textit{V. Patkar} [Z. Angew. Math. Mech. 64, No. 1, 59-62 (1984; Zbl 0537.90089)].