Automated Parameterization of Performance Models from Measurements

Estimating parameters of performance models from empiri- cal measurements is a critical task, which often has a major in uence on the predictive accuracy of a model. This tuto- rial presents the problem of parameter estimation in queue- ing systems and queueing networks. The focus is on reliable estimation of the arrival rates of the requests and of the ser- vice demands they place at the servers. The tutorial covers common estimation techniques such as regression methods, maximum-likelihood estimation, and moment-matching, dis- cussing their sensitivity with respect to data and model char- acteristics. The tutorial also demonstrates the automated estimation of model parameters using new open source to