|
Activity Number:
|
203
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 30, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Physical and Engineering Sciences
|
| Abstract - #310287 |
|
Title:
|
Comparison Sequential Testing for Reliability: Optimal Truncation of Short Tests
|
|
Author(s):
|
Yefim H. Michlin*+ and Genady Grabarnik and Elena Leshchenko
|
|
Companies:
|
Technion and IBM T.J. Watson Research Center and Technion
|
|
Address:
|
Quality Assurance and Reliability , Haifa , 32000, Israel
|
|
Keywords:
|
Sequential testing ; Comparison testing ; MTBF ratio ; Sample number
|
|
Abstract:
|
The study deals with simultaneous testing of two systems, one "basic" (b) and the other "new" (n) - with an exponential distribution of their times between failures. It is checked whether the mean TBFn/MTBFb ratio (discrimination ratio - DR) equals a given value, versus whether it is smaller than the latter. Results are reported for tests of this type with large DR, i.e. with small average sample numbers (ASN). Such tests serve for fast comparison of real-life systems at early stages of development. In this paper it is established that the distribution density of possible pairs of first and second type errors decreases steeply with decrease of the ASN and with test truncation. This gives rise to difficulties in optimizing the solution. An algorithm and methodology are presented, yielding an effective assessment of test optimality and planning in fullest compliance with requirements.
|
