DIRECT REDUCTION OF LARGE CONCAVE PROGRAMS
Abstract
A very simple and flexible procedure is given for reducing a concave mathematical program to a finite sequence of subproblems, each involving a subset of the original constraints. While very much in the spirit of the 'secondary constraint' idea, constraints are dropped as well as added from subproblem to subproblem. The reduction procedure is of quite general applicability, and is designed to be used with various concave programming algorithms for solving the subproblems. It seems particularly suited to facilitating the solution of problems with more constraints than could otherwise be handled. It is shown that an obvious specialization to linear programs yields the dual Simplex method, and that various 'cutting-plane' methods can also be naturally described from the general viewpoint offered in the report.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1966
- Accession Number
- AD0647552
Entities
People
- A. M. Geoffrion
Organizations
- University of California, Los Angeles