Abstract
In this paper we study very large-scale neighborhoods for the minimum total weighted completion time problem on parallel machines, which is known to be strongly NP-hard. We develop two different ideas leading to very large-scale neighborhoods in which the best improving neighbor can be determined by calculating a weighted matching. The first neighborhood is introduced in a general fashion using combined operations of a basic neighborhood. Several examples for basic neighborhoods are given. The second approach is based on a partitioning of the job sets on the machines and a reassignment of them. In a computational study we evaluate the possibilities and the limitations of the presented very large-scale neighborhoods.
Original language | Undefined |
---|---|
Pages (from-to) | 637-658 |
Number of pages | 22 |
Journal | Journal of heuristics |
Volume | 17 |
Issue number | 6 |
DOIs | |
Publication status | Published - Dec 2011 |
Keywords
- MSC-68M20
- MSC-90B35
- IR-78565
- Scheduling · Parallel machines · Total weighted completion time · Verylarge-scale neighborhoods · Local search
- METIS-281581
- EWI-20836