Confidence Intervals for the Ratio of Two Means with Application to Simulations.
CONTROL ANALYSIS CORP PALO ALTO CALIF
Pagination or Media Count:
In simulating stable stochastic systems one is frequently confronted by the need to produce confidence intervals for the ratio of the expected values of two random variables r.v.s. This situation arises, for example, when one simulates the general many-server queue or stochastic systems modeled as Markov chains in either discrete or continuous time. The report derives a confidence interval for the ratio of two means which is a distinct improvement over an old method. Both the old and new confidence intervals are illustrated by numerical examples. Author
- Statistics and Probability
- Operations Research