The Accelerated Bound-and-Scan Algorithm for Integer Programming
Abstract
This paper presents a new implicit enumeration algorithm for solving the pure integer linear programming problem. The theory of equivalent integer programming problems is first used to reformulate the problem. A technique, originally used with particular success in the bound-and-scan algorithm to deal with only a subset of the variables, is extended to all of the variables in the restructured problem. In addition to the resulting basic enumeration scheme, the algorithm includes a scanning procedure and a method for identifying constraints which become redundant during the course of the algorithm. Computational experience on standard test problems is reported.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 15, 1972
- Accession Number
- AD0743128
Entities
People
- Bruce H. Faaland
- Frederick S. Hiller
Organizations
- Stanford University