arXiv Analytics

Sign in

arXiv:1512.07771 [math.PR]AbstractReferencesReviewsResources

Achievable Performance of Blind Policies in Heavy Traffic

Nikhil Bansal, Bart Kamphorst, Bert Zwart

Published 2015-12-24Version 1

For a GI/GI/1 queue, we show that the average sojourn time under the (blind) Randomized Multilevel Feedback algorithm is no worse than that under Shortest Remaining Processing Time algorithm times a logarithmic function of the system load. Moreover, it is verified that this bound is tight in heavy traffic, up to a constant multiplicative factor. We obtain this result by combining techniques from two disparate areas: competitive analysis and applied probability.

Related articles: Most relevant | Search more
arXiv:1104.0167 [math.PR] (Published 2011-04-01, updated 2011-12-06)
Gaussian queues in light and heavy traffic
arXiv:1607.01744 [math.PR] (Published 2016-07-06)
Diffusion Approximations for Double-ended Queues with Reneging in Heavy Traffic
arXiv:0807.4621 [math.PR] (Published 2008-07-29, updated 2014-05-31)
On the $M_t/M_t/K_t+M_t$ queue in heavy traffic