Competitive ratio of List Scheduling on uniform machines and randomized heuristics
- 130 Downloads
We study online scheduling on m uniform machines, where m−1 of them have a reference speed 1 and the last one a speed s with 0≤s≤1. The competitive ratio of the well-known List Scheduling (LS) algorithm is determined. For some values of s and m=3, LS is proven to be the best deterministic algorithm. We describe a randomized heuristic for m machines. Finally, for the case m=3, we develop and analyze the competitive ratio of a randomized algorithm which outperforms LS for any s.
KeywordsAnalysis of algorithms Online algorithms Competitive ratio Randomized scheduling
Unable to display preview. Download preview PDF.
- Hochbaum, D. S. (1997). Approximation algorithms for NP-hard problems. Boston: PWS Publishing Company. Google Scholar
- Kokash, N. (2004). An efficient heuristic for on-line scheduling in system with one fast machine. Master Thesis. Google Scholar
- Sgall, J. (1994). On-line scheduling on parallel machines. Ph.D. Thesis, Carnegie-Mellon University, Pittsburgh, PA. Google Scholar