Some Applications of Penalty Functions in Mathematical Programming.
Abstract
Penalty function minimization is a useful technique for converting constrained optimization problems to simpler unconstrained optimization problems. One difficulty with this approach has been the determination of the size of an adequate penalty parameter. This work shows how to choose precisely the penalty parameter in order to meet any preassigned accuracy. In addition penalty functions are used to obtain bounds on the size of a solution of a constrained optimization problem without solving it. The authors also shows how his results can be used to solve huge sparse linear programs to any desired degree of accuracy.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 01, 1984
- Accession Number
- ADA147443
Entities
People
- Olvi L. Mangasarian
Organizations
- University of Wisconsin–Madison