ISSN:
1436-4646
Source:
Springer Online Journal Archives 1860-2000
Topics:
Computer Science
,
Mathematics
Notes:
Abstract In this study we introduce the notion of supporting functions and exploit some of their important properties. We make use of these functions to develop generalized optimality conditions of a mixed stationary-saddle type, where the supporting functions play either the role of the gradients/directional derivatives and/or the role of the original functions. These conditions may be applied to certain problems involving nondifferentiable and/or nonconvex functions. Classical, as well as new stationary and saddle optimality conditions follow from our approach in a natural fashion.
Type of Medium:
Electronic Resource
URL:
http://dx.doi.org/10.1007/BF01580133