DUALITY IN NONLINEAR PROGRAMMING: A SIMPLIFIED APPLICATIONS-ORIENTED DEVELOPMENT,
Abstract
A simpler, more versatile, and more general approach is presented to the computational solution of complex practical optimization problems of nonlinear programming through their duals. Duality theory gives 'prices' for scarce resources, and certain planning problems are more easily solved by passing to an equivalent dual problem. However, the number of applications so far is small compared to the number of theoretical papers. This study avoids differential calculus, minimax theory, and the conjugate functions or fixed-point theorems employed by most studies in duality theory. Instead, it uses only assumptions that are likely to be verifiable, and relies on the relatively elementary theory of convexity and the concept of a perturbation function--the optimal value of a program as a function of perturbations of its right-hand side. The approach gives useful geometric and mathematical insights and is closely interwoven with optimality theory. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1970
- Accession Number
- AD0709184
Entities
People
- A. M. Geoffrion
Organizations
- RAND Corporation