Abstract
In this paper, we study 0–1 quadratic programs with joint probabilistic constraints. The row vectors of the constraint matrix are assumed to be normally distributed but are not supposed to be independent. We propose a mixed integer linear reformulation and provide an efficient semidefinite relaxation of the original problem. The dependence of the random vectors is handled by the means of copulas. Finally, numerical experiments are conducted to show the strength of our approach.
Original language | English (US) |
---|---|
Pages (from-to) | 1283-1295 |
Number of pages | 13 |
Journal | Optimization Letters |
Volume | 9 |
Issue number | 7 |
DOIs | |
State | Published - Oct 22 2015 |
Externally published | Yes |
Keywords
- 0–1 quadratic program
- Copula theory
- Joint probabilistic constraints
- Semidefinite programming
- Stochastic programming
ASJC Scopus subject areas
- Control and Optimization