Sankar Basu
Journal of the Franklin Institute
We study an M/M/1 queueing system under the shortest remaining processing time (SRPT) policy. We show that the average sojourn time varies as Θ((μ(1-ρ) ln(e/(1-ρ)))-1), where ρ is the system load. Thus, SRPT offers a Θ(ln(e/(1-ρ))) factor improvement over policies that ignore knowledge of job sizes while scheduling. © 2004 Elsevier B.V. All rights reserved.
Sankar Basu
Journal of the Franklin Institute
Zhihua Xiong, Yixin Xu, et al.
International Journal of Modelling, Identification and Control
Heng Cao, Haifeng Xi, et al.
WSC 2003
Nikhil Bansal, Rohit Khandekar, et al.
SIAM Journal on Computing