W e develop a numerically exact method for evaluating the time-dependent mean, variance, and higher-order moments of the number of entities in a Ph t /Ph t / queueing system. We also develop a numerically exact method for evaluating the distribution function and moments of the virtual sojourn time for any time t; in our setting, the virtual sojourn time is… (More)
A statistically efficient method for performing simulation experimentation of nonstationary queueing models is outlined. The method utilizes a nonstationary queueing approximation as an external control variate system. A simple nonstationary tandem queueing model serves as an example of this variance reduction method.
Moment-differential equations, moment-matching, surrogate distributions and numerical integration have been combined to form a small closed set of differential equations that can accurately approximate huge sets of Kolmogorov-forward equations for complex nonstationary queueing models.