Computer Science Department
School of Computer Science, Carnegie Mellon University
The Effect of Higher Moments of Job Size Distribution
Varun Gupta, Jim Dai*, Mor Harchol-Balter, Bert Zwart*
The M/G/K queueing system is the oldest model for multi-server systems, and has been the topic of performance papers for almost half a century. However, even now, only coarse approximations exist for its mean waiting time. All the closed-form (non-numerical) approximations in the literature are based on the first two moments of the job size distribution. In this paper we prove that no approximation based on only the first two moments can be accurate for all job size distributions, and we provide a lower bound on the inapproximability ratio. This is the first such result in the literature. The proof technique behind this result is novel as well and combines mean value analysis, sample path techniques, scheduling, regenerative arguments, and asymptotic estimates. Finally, our work provides insight into the effect of higher moments of the job size distribution on the mean waiting time.
*School of Industrial and Systems Engineering, George Institute of Technology, Atlanta, GA