Abstract
We discuss relationships between the solution to an integer-programming problem and the solution to its relaxed linear-program-ming problem in terms of the distance that separates them and related bounds. Assuming knowledge of a subset of extreme points, we develop bounds for associated convex combinations and show how improved bounds on the integer-programming problem's objective function and variables can be obtained.
Original language | English (US) |
---|---|
Pages (from-to) | 55-63 |
Number of pages | 9 |
Journal | Journal of Optimization Theory and Applications |
Volume | 98 |
Issue number | 1 |
DOIs | |
State | Published - Jul 1998 |
Keywords
- Extreme points
- Integer programming
- Relaxed linear-programming problems
- Solution bounds
ASJC Scopus subject areas
- Control and Optimization
- Management Science and Operations Research
- Applied Mathematics