Average regression-adjusted controlled regenerative estimates

Computer simulations of queuing systems are often used to generate estimates of system characteristics along with estimates of their precision. It is pointed out that obtaining precise estimates, especially for high traffic intensities, can require large amounts of computer time. Average regression-adjusted controlled regenerative estimates result from combining the two techniques of controlled regenerative estimates and average regression-adjusted regenerative estimates. Combining these two techniques can create estimates whose estimated mean-square error is much lower than can be obtained through using either technique alone. An example using data from a simulation of an M/M/1 queue is provided. In this example the estimated mean square error for the average regression-adjusted controlled regenerative estimate is just 10% of the mean square error estimate for the straightforward regenerative estimate