Online Matching for Scheduling Problems
In this work an alternative online variant of the matching problem in bipartite graphs is presented. It is motivated by a scheduling problem in an online environment. In such an environment, a task is unknown up to its disclosure. However, in that moment it is not necessary to take a decision on the service of that particular task. In reality, an online scheduler has to decide on how to use the current resources. Therefore, our problem is called online request server matching (ORSM). It differs substantially from the online bipartite matching problem of Karp et al. [KVV90]. Hence, the analysis of an optimal, deterministic online algorithm for the ORSM problem results in a smaller competitive ratio of 1.5.
We also introduce an extension to a weighted bipartite matching problem. A lower bound of √+1 /2 ≈ 1.618 and an upper bound of 2 is given for the competitive ratio.
KeywordsCompetitive Ratio Online Algorithm Memory Module Online Schedule Bipartite Match
Unable to display preview. Download preview PDF.
- [BEY98]Allan Borodin and Ran El-Yaniv. Online Computation and Competitive Analysis. Cambridge University Press, 1998.Google Scholar
- [BR93]Ethan Bernstein and Sridhar Rajagopalan. The roommates problem: Online matching on general graphs. Technical Report CSD-93-757, University of California, Berkeley, CS Department, 1993.Google Scholar
- [KVV90]Richard M. Karp, Umesh V. Vazirani, and Vijay V. Vazirani. An optimal algorithm for on-line bipartite matching. In Proceedings of the 22nd Annual ACM Symposium on Theory of Computing, pages 352–358, New York, 1990. ACM Press.Google Scholar
- [Rie98]Marco Riedel. Online request server matching. Technical Report tr-ri-98-195, University of Paderborn, Germany, April 1998. available via: ftp://ftp.uni-paderborn.de/doc/techreports/Informatik/tr-ri-98-195.ps.Z Google Scholar