US Army Materiel Command AMCP 706-197 — AMCP 706-197 Reliability Prediction Handbook — US Army Engineering Design Guide PDF

Downloaded from https://www.everyspec.com AMCP 706.197 2-5 DISTRIBUTIONS the integers, it is convenient to introduce the notion of random variable. For :'xample, the Very often the sample space is a subset of events Cj and E, in this chapter are random the integers (or can be put into 1 - 1 corre- variables, and the probability of the event spondence with some of the integers), and the depends on the integer i. A variable is a ran- probability to be assigned to a sample point is dom variable if the uncertainty involved with a function of the integer which corresponds it is important, i.e., if probabilities need to be to the sample point. The probability mass associated with it.

This is an engineering function (pmf) is the function which assigns a decision; for example, the lengths of posts to probability to each sample point. This is illus- be driven in the ground might not be con- trated in Table 2-5. sidered random even though they had a spread of ± 10%, whercas the diameters of 2-5.1 RANDOM VARIABLES ball bearings would probably be random vari- ables if their spread was ± 1%. When the sample space is associated with TABLE 2-5. DISCRETE DISTRIBUTIONS Linomial Poisson ° parameters PI ,P 2 ,N . (P 1 + P 2 = 1) random variables nj, n 2 n (n + n 2 =N) pmf N! PiP2 M j" PI P 2 n, !n 2 !

nI mean p pIN,P 2 N variance 02 PIP 2 N I 3rd central moment M 3 NP l P 2 (P 2 - ) / 0 (P2 ) coefficient of variation - 4th central moment M 4 NpIp 2 (3NpIp 2 -6PiP 2 + 1) /'( 3 /u + 1) coefficiert of skewness M 3 P 2 - P' (NpIP 2 )% excess coefficient of kurtosis M 4 6 1 a4N NpIp 2 *As is customary, the symbol p (for mean) is used for the parameter because the param- eter happens to be the mean, This is also done in the s-normal distribution. 2-10