Nonparametric selection procedures applied to state traffic fatality rates
WSC '77 Proceedings of the 9th conference on Winter simulation - Volume 1
New procedures for selection among (simulated) alternatives
WSC '77 Proceedings of the 9th conference on Winter simulation - Volume 1
| Hi-index | 0.00 |
When k alternatives are simulated (e.g. in a study to find which one optimizes the performance of some real or planned system) it is not rare for multiple criteria to be employed (i.e., several system responses are of interest and no simple combination of them, called a measure of effectiveness, is available). In such settings, n persons (who are responsible for the system, e.g. administrators, investigators, etc.) may be asked to rank the performances of the k from best to worst. An optimal method for performing this ranking will be discussed. A similar problem occurs when each of n referees ranks k contestants to select the winner of a prize (the best contestant), and it is common that not all the referees complete the full ranking. Similar situations arise when selecting the most efficient simulation algorithm. Our methods for selecting the winner also apply to data of this type.