A continuous approach for the concave cost supply problem via DC programming and DCA

The paper addresses an important but difficult class of concave cost supply management problems which consist in minimizing a separable increasing concave objective function subject to linear and disjunctive constraints. We first recast these problems into mixed zero-one nondifferentiable concave mi...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Discrete Applied Mathematics Ročník 156; číslo 3; s. 325 - 338
Hlavní autoři: Le Thi, Hoai An, Pham Dinh, Tao
Médium: Journal Article Konferenční příspěvek
Jazyk:angličtina
Vydáno: Lausanne Elsevier B.V 01.02.2008
Amsterdam Elsevier
New York, NY
Témata:
Algorithmics. Computability. Computer arithmetics
Applied sciences
Combinatorics
Combinatorics. Ordered structures
Computer Science
Computer science; control theory; systems
Concave cost supply problem
Concave programming
DC (Difference of Convex functions) programming
DCA
Disjunctive constraints
Exact penalty
Exact sciences and technology
Mathematics
Mixed 0-1 concave programming
Reformulation
Sciences and techniques of general use
Theoretical computing
Concave programming
Exact penalty
DCA
DC (Difference of Convex functions) programming
Reformulation
Disjunctive constraints
Concave cost supply problem
Mixed 0-1 concave programming
Costs
Optimization method
Convex programming
Penalty method
Efficiency
Increasing function
Convex function
Robustness
Combinatorics
Deterministic approach
Computer theory
Minimization
Management
Algorithm
Constrained optimization
Concave function
Numerical simulation
Mixed problem
Objective function
ISSN:0166-218X, 1872-6771
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:The paper addresses an important but difficult class of concave cost supply management problems which consist in minimizing a separable increasing concave objective function subject to linear and disjunctive constraints. We first recast these problems into mixed zero-one nondifferentiable concave minimization over linear constraints problems and then apply exact penalty techniques to state equivalent nondifferentiable polyhedral DC (Difference of Convex functions) programs. A new deterministic approach based on DC programming and DCA (DC Algorithms) is investigated to solve the latter ones. Finally numerical simulations are reported which show the efficiency, the robustness and the globality of our approach.
ISSN:0166-218X
1872-6771
DOI:10.1016/j.dam.2007.03.024