[EFS00] Decomposition and Linearization for 0-1 Quadratic Programming
Revue Internationale avec comité de lecture :
Journal Annals of Operations Research,
vol. 99(1), 2000
motcle:
Résumé:
This paper presents a general decomposition method to compute bounds for constrained 0-1 quadratic programming. The best decomposition is found by using a Lagrangean decomposition of the problem. Moreover, in its simplest version this method is proved to give at least the bound obtained by the LP-relaxation of a non trivial linearization. To illustrate this point, some computational results are given for the 0-1 quadratic knapsack problem.
Commentaires:
pp. 79-93