Optimality criteria for general unconstrained geometric programming problems

The paper presents a generalization of geometric programming problems based on a new definition of the conjugate function. This results in a primal-dual pair of geometric programs which are symmetric in the sense that the dual objective function is independent from the set of primal feasible solutions (new result) and the primal objective function is independent from the set of dual feasible solutions (usual result). The definition of conjugate functions used to achieve this result is: Let a function g: \(D_ g\to R\) (where \(D_ g\subset R^ n\) is the domain of g) be defined. Denote h(y)\(\equiv \sup_{x\in D_ g}\{g(x)-| <x,y>\}\), where h: \(R^ n\to R\) is the conjugate function of g defined on the set \(D_ h\equiv \{y\in R^ n:\sup_{x\in D_ g}\{g(x)- <x,y>\}<\infty \}\subset R^ n.\) The authors show that the basic optimality criteria hold under this new generalization.

Recommendations

Cites work

Cited in

(4)

This page was built for publication: Optimality criteria for general unconstrained geometric programming problems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q811403)