Initial transient problem for steady-state output analysis
WSC '05 Proceedings of the 37th conference on Winter simulation
Hi-index | 0.00 |
In this paper, Deutsch, Richards and Fernandez-Torres (1983) fit relaxed time series (RARMA) models to transient and steady state data from a simulated M/M/1 queue and a nonstationary queueing network. Both systems are initially empty and idle. The statistical identification procedure (Richards 1983) for the RARMA model class is illustrated. Improved estimates of the steady state mean of the M/M/1, and of the polynominal in time of the network's nonstationary “steady state” performance, are obtained by including the identified transient model in the RARMA regression.