Method of determining whether an improved item has a better mean lifetime than an existing item
AssigneeThe United States of America as represented by the Secretary of the Army
InventorsAndrew G. Glen, Bobbie L. Foote
Life of this patent
Estimating the term…
Abstract
A method of determining whether an improved item has a better mean lifetime than an existing item having a fully specified lifetime distribution function. A finite number n of examples of the new item are placed on life test, and the failure times of the n examples of the new item are recorded as they occur. At each occurrence of failure, the newly noted failure time is entered into a list of previously noted failure times and the most up-to-date statistical P-value is calculated for the life test so far. When the statistical P-value is sufficiently small there is statistical inference that the new item is better than the current one, and the test is ended.
Claims1
A method of determining whether an improved item has a better mean lifetime than an existing item with a fully specified lifetime distribution function, comprising the steps of:
- (a) placing a finite number n examples of the improved item on life test simultaneously with the intent of letting some or all of the n examples fail;
- (b) observing the first failure time;
- (c) computing the first statistical P-value;
- (d) if the first statistical P-value is less than 0.05, ending the test early;
- (e) if the test is not ended, observing the next failure time;
- (f) computing the next statistical P-value;
- (g) if the next statistical P-value is less than 0.05, ending the test early;
- (h) if the test is not ended, asking whether the last example has failed;
- (i) if the last example has not failed, repeating steps (e)–(i);
- and
- (j) if the last example has failed, ending the test.
Dependencies are inferred from the claim text.