Optimality conditions and DC-Dinkelbach-type algorithm for generalized fractional programs with ratios of difference of convex functions

In this paper, we develop optimality conditions and propose an algorithm for generalized fractional programming problems whose objective function is the maximum of finite ratios of difference of convex (dc) functions, with dc constraints, that we will call later, DC-GFP. Such problems are generally...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Optimization letters Ročník 15; číslo 7; s. 2351 - 2375
Hlavní autoři: Ghazi, Abdelouafi, Roubi, Ahmed
Médium: Journal Article
Jazyk:angličtina
Vydáno: Berlin/Heidelberg Springer Berlin Heidelberg 01.10.2021
Témata:
Computational Intelligence
Mathematics
Mathematics and Statistics
Numerical and Computational Physics
Operations Research/Decision Theory
Optimization
Original Paper
Simulation
DC programming
Optimality conditions
Fractional programming
Convex programming
Dinkelbach algorithms
ISSN:1862-4472, 1862-4480
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:In this paper, we develop optimality conditions and propose an algorithm for generalized fractional programming problems whose objective function is the maximum of finite ratios of difference of convex (dc) functions, with dc constraints, that we will call later, DC-GFP. Such problems are generally nonsmooth and nonconvex. We first give in this work, optimality conditions for such problems, by means of convex analysis tools. For solving DC-GFP, the use of Dinkelbach-type algorithms conducts to nonconvex subproblems, which causes the failure of the latter since it requires finding a global minimum for these subprograms. To overcome this difficulty, we propose a DC-Dinkelbach-type algorithm in which we overestimate the objective function in these subproblems by a convex function, and the constraints set by an inner convex subset of the latter, which leads to convex subproblems. We show that every cluster point of the sequence of optimal solutions of these subproblems satisfies necessary optimality conditions of KKT type. Finally we end with some numerical tests to illustrate the behavior of our algorithm.
ISSN:1862-4472
1862-4480
DOI:10.1007/s11590-020-01694-w