Citation
Abstract
The problem is to test whether the frequency of random events (e.g., DSIF equipment failures) is at a nominally prescribed value. When the actual frequency is higher, a determination of this fact is to be made as quickly as possible. A test based on sequential maximum likelihood ratio methods is developed and approximations of its performance characteristics are derived. Results of Monte Carlo sampling demonstrate that these approximations are accurate and that high statistical efficiency is attained over a broad range of possible higher frequencies. Some applications to reliability and inventory policies for the DSIF are indicated.
Details
- Volume
- V
- Published
- October 15, 1971
- Pages
- 82–90
- File Size
- 870.1 KB