A differentiable dual approach to large scale 01-problems
We present a differentiable dual approach to binary optimization problems. We solve a sequence of, parameterized, differentiable dual problems, which in the limit converge to the standard Lagrangean dual. We start with a large parameter value, giving us a very smooth dual problem. For a given parameter value the dual is maximized by a coordinate ascent method. The parameter is successively decreased towards zero, yielding, in the limit, a good estimate of the optimum value of the LP-relaxation. In addition, the relaxed primal solutions, interpreted as probabilities, enable us to generate near optimal primal feasible solutions. The method is applied to a set of set-covering problems originating from airline crew scheduling.
- OSTI ID:
- 36344
- Report Number(s):
- CONF-9408161-; TRN: 94:009753-0684
- Resource Relation:
- Conference: 15. international symposium on mathematical programming, Ann Arbor, MI (United States), 15-19 Aug 1994; Other Information: PBD: 1994; Related Information: Is Part Of Mathematical programming: State of the art 1994; Birge, J.R.; Murty, K.G. [eds.]; PB: 312 p.
- Country of Publication:
- United States
- Language:
- English
Similar Records
LP and NLP decomposition without a master problem
Running Primal-Dual Gradient Method for Time-Varying Nonconvex Problems