Abstract
We explore in this paper certain rich geometric properties hidden behind quadratic 0-1 programming. Especially, we derive new lower bounding methods and variable fixation techniques for quadratic 0-1 optimization problems by investigating geometric features of the ellipse contour of a (perturbed) convex quadratic function. These findings further lead to some new optimality conditions for quadratic 0-1 programming. Integrating these novel solution schemes into a proposed solution algorithm of a branch-and-bound type, we obtain promising preliminary computational results.
| Original language | English |
|---|---|
| Pages (from-to) | 797-829 |
| Journal | Journal of Global Optimization |
| Volume | 52 |
| Issue number | 4 |
| Online published | 9 Apr 2011 |
| DOIs | |
| Publication status | Published - Apr 2012 |
| Externally published | Yes |
Research Keywords
- Branch-and-bound method
- Lower bounds
- Nonlinear integer programming
- Optimality condition
- Quadratic 0-1 programming
- Variable fixation
Fingerprint
Dive into the research topics of 'An exact solution method for unconstrained quadratic 0-1 programming: a geometric approach'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver