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 CHAPTER 2- REVIEW OF ELEMENTARY PROBABILITY THEORY (DISCRETE) 2-0 LIST OF SYMBOLS particular. Statistics on the other hand treats A,B,C,E = sets actual data and tries to decide what useful AB, =vthings can be done with them and how to got A,,AB,B 0 = evens that units o and them. It goes from the particular to the gener- U 8 = s Failed or Good al. A statistic is a number obtained from a Ea ,E HT,EET = events of Benign, High Tern- statistics.

In engineering problems one usually E{}= -expected vhlue of sample or obtained from manipulating other Electrical - uses a mixture of probability and statistics; senturoe nthere is little to be gained in debating which sient environments calculations are probabilistic and which are EL,Es = events of Light and Severe statistical. environments M, ith central moment 2-2 BASIC PROBABILITY RULES NA ,N B JE number of subsets in A,B,E pmf probability mass function 2-2.1 SAMPLE SPACE, SAMPLE POINT, Pr{ } = probability of EVENT s- = denotes statistical definition 2 = standard deviation bility problem. The sample space is made up Svariance of all the sample points.

An event is a collec- = compleie sample space tion of sample points; it can contain as few f= null event sample points as none, or as many as all. The u = union concepts are best illustrated by examples. See n = intersection the Bibliography in Chapter 1 for books 2-1 INTRODUCTION which can explain the concepts. ff= mean These are basic concepts for any proba- Example 1. For one throw of a single die, Tthe sample space is the set of numbers 1, 2, 3, ~The question always arises "What is prob- ability?" Some say it is relative frequency; 4, 5, 6; i.e., the sample space is all possible values that man arise.

Each value is called a Sothers say it is degree-of-belief; and still others have different concepts. In many good sample poir:. There are six sampe points in ~the samph space for this example. reliability and engineering textbooks (and virtually all mathematical books) probabilities Every possible single outcome of an ex- are mathematical concepts which can then be periment is a sample point. The naming of appliod to such things as relative frequency every sample point is a first step in making a and degree-of-belier. The situation is analo- probabilistic nodel of any problem, although gous to plane geometry. Plane geometry is a it often is done implicitly.

Each sample point mathematical theory that uses concepts such also has a proba'ility associated with it. The as point and line. The theory is true (consist- probability usually is assigned or calculated ent) regardless of what a point or line is taken from known event-probabilities. to be. Plane geometry often is applied success- In the example of one throw of a single fully to many reasonably flat things in every- die, the probabilities usually are assigned by day life, and we associate point and line with defining the die to be "fair"; i.e., each face the everyday concepts. has an equal probability of appearing.

Then Probability and statistics are related very the probability assigned to each sample point closely to each other. The difference between is 1/6. By definition, the sum of the proba- them is not clear to many engineers. Proba- hilities for all sample points must be one. bility theory usually considers the parameters Engineers who use probability often go of a general problem as known, then com- astray because they do not understand sam- putes numbers (probabilities) about particular pie-space and assignment of probabilities to 'ets of events. It goes from the general to the each sample point. 2-1