Springe zum Hauptinhalt
Fakultät für Mathematik
Fakultät für Mathematik
K.Beer; G.Zenker : The marginal value in quadratic programming with unbounded solution sets

K.Beer; G.Zenker : The marginal value in quadratic programming with unbounded solution sets


Author(s) :
K.Beer; G.Zenker
Title :
The marginal value in quadratic programming with unbounded solution sets
Preprint series
Technische Universität Chemnitz, Fakultät für Mathematik (Germany). Preprint 98-23, 1998
Mathematics Subject Classification :
90C31 [ Sensitivity, etc. ]
90C25 [ Convex programming ]
Abstract :
A formula for computing the diretional derivative of the optimal value function of a convex quadratic programming problem is given. Pertubations in the constrained matrix are not allowed. For this situation three equvalent necessary and sufficient criteria for the existence of the directional derivative together with an upper and lower bound for the expansion of optimal value are proved.
Keywords :
directional derivative, marginal function, quadratic programming, parametric optimization, nondifferentiable optimization, sensitivity analysis
Language :
english
Publication time :
10/1998