Knapsack problems with dependencies through non-additive measures and Choquet integral

In portfolio selection problems the items often depend on each other, and their synergies and redundancies need to be taken into account. We consider the knapsack problem in which the objective is modelled as the Choquet integral with respect to a supermodular capacity which quantifies possible syne...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:European journal of operational research Ročník 301; číslo 1; s. 277 - 286
Hlavní autor: Beliakov, Gleb
Médium: Journal Article
Jazyk:angličtina
Vydáno: Elsevier B.V 16.08.2022
Témata:
Capacities
Choquet integral
Fuzzy measures
Fuzzy sets
Integer programming
Optimisation
Shapley value
Integer programming
Fuzzy sets
Fuzzy measures
Shapley value
Choquet integral
Optimisation
Capacities
ISSN:0377-2217, 1872-6860
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:In portfolio selection problems the items often depend on each other, and their synergies and redundancies need to be taken into account. We consider the knapsack problem in which the objective is modelled as the Choquet integral with respect to a supermodular capacity which quantifies possible synergies. We provide various formulations which lead to the standard linear mixed integer programs, applicable to small and large portfolios. We also study scalability of the solution methods and compare large problems defined with respect to 2-additive capacities which model pairwise interactions, and linear knapsack with respect to the Shapley values of these capacities.
ISSN:0377-2217
1872-6860
DOI:10.1016/j.ejor.2021.11.004