Semi-infinite fractional programming (Q2406457)

As its title indicates, this book deals with constrained optimization problems in which the objective function is a fraction (or the pointwise maximum of a collection of fractions) and either the number of variables or the number of constraints is infinite, while the other number is finite. Such problems are analyzed under the assumption that the functions involved satisfy suitable generalized invexity conditions. The book consists of the following twelve chapters: 1. Higher order parametric duality models. 2. New generation parametric optimality. 3. Accelerated roles for parametric optimality. 4. Semi-infinite multiobjective fractional programming I. 5. Semi-infinite multiobjective fractional programming II. 6. Semi-infinite multiobjective fractional programming III. 7. Hanson-Antczak-type generalized V-invexity I. 8. Parametric optimality in semi-infinite fractional programs. 9. Semi-infinite discrete minmax fractional programs. 10. Next-generation semi-infinite discrete fractional programs. 11. Hanson-Antczak-type sonvexity III. 12. Semi-infinite multiobjective optimization. The following definition may be useful to get a flavor of the generalized invexity notions used in this book: a twice differentiable real-valued function \(f\) defined on an open subset of \(\mathbb{R}^{n}\) is said to be \( (\phi ,\eta ,\zeta ,\rho ,\theta ,m)\)-sonvex at \(x^{\ast }\) if for each \( x\neq x^{\ast }\) and \(z\in \mathbb{R}^{n}\), one has \(\phi (f(x)-f(x^{\ast }))\geq \left\langle \nabla f(x^{\ast }),\eta (x,x^{\ast })\right\rangle + \frac{1}{2}\left\langle \zeta (x,x^{\ast }),\nabla ^{2}f(x^{\ast })z\right\rangle +\rho (x,x^{\ast })\left\| \theta (x,x^{\ast })\right\| ^{m}\). Here \(\phi \), \(\eta \), \(\zeta \), \(\rho \) and \(\theta \) are functions, \(m\) is a positive integer, and \(\left\| \cdot \right\| \) is a norm. The word ``sonvex'' is an abbreviation of ``second-order invex.'' Other generalized invexity notions involving Hadamard directional derivatives are also considered.