PROBABILITY THEORY:CONTINUOUS PROBABILITY DISTRIBUTIONS

By -
CONTINUOUS PROBABILITY DISTRIBUTIONS

Section 24.2 defined continuous random variables as those random variables that take on a continuum of values. The CDF for a continuous random variable FX(b) can usually be written as

INTRODUCTION TO OPERATIONS RESEARCH-0557

where fX(y) is known as the density function of the random variable X. From a notational standpoint, the subscript X is used to indicate the random variable that is under consideration. When there is no ambiguity, this subscript may be deleted, and fX(y) will be de- noted by f(y). It is evident that the CDF can be obtained if the density function is known. Furthermore, a knowledge of the density function enables one to calculate all sorts of probabilities, for example,

INTRODUCTION TO OPERATIONS RESEARCH-0558

Note that strictly speaking the symbol P{a < X ::: b} relates to the probability that the outcome w of the experiment belongs to a particular event in the sample space, namely, that event such that X(w) is between a and b whenever w belongs to the event. However, the reference to the event will be suppressed, and the symbol P will be used to refer to the probability that X falls between a and b. It becomes evident from the previous ex- pression for P{a < X ::: b} that this probability can be evaluated by obtaining the area under the density function between a and b, as illustrated by the shaded area under the density function shown in Fig. 24.8. Finally, if the density function is known, it will be said that the probability distribution of the random variable is determined.

Naturally, the density function can be obtained from the CDF by using the relation

INTRODUCTION TO OPERATIONS RESEARCH-0559

For a given value c, P{X = c} has not been defined in terms of the density function. However, because probability has been interpreted as an area under the density function, P{X = c} will be taken to be zero for all values of c. Having P{X = c} = 0 does not mean that the appropriate event E in the sample space (E contains those w such that X(w) = c) is an impossible event. Rather, the event E can occur, but it occurs with probability zero. Since X is a continuous random variable, it takes on a continuum of possible values, so that selecting correctly the actual outcome before experimentation would be rather star- tling. Nevertheless, some outcome is obtained, so that it is not unreasonable to assume that the preselected outcome has probability zero of occurring. It then follows from P{X = c} being equal to zero for all values c that for continuous random variables, and any a and b,

INTRODUCTION TO OPERATIONS RESEARCH-0560

In defining the CDF for continuous random variables, it was implied that fX(y) was defined for values of y from minus infinity to plus infinity because

INTRODUCTION TO OPERATIONS RESEARCH-0561

INTRODUCTION TO OPERATIONS RESEARCH-0562

The exponential distribution has had widespread use in operations research. The time between customer arrivals, the length of time of telephone conversations, and the life of electronic components are often assumed to have an exponential distribution. Such an assumption has the important implication that the random variable does not “age.” For example, suppose that the life of a vacuum tube is assumed to have an ex- ponential distribution. If the tube has lasted 1,000 hours, the probability of lasting an additional 50 hours is the same as the probability of lasting an additional 50 hours, given that the tube has lasted 2,000 hours. In other words, a brand new tube is no “better” than one that has lasted 1,000 hours. This implication of the exponential distribution is quite important and is often overlooked in practice.

The Gamma Distribution

A continuous random variable whose density is given by

INTRODUCTION TO OPERATIONS RESEARCH-0563

A graph of a typical gamma density function is given in Fig. 24.11.

A random variable having a gamma density is useful in its own right as a mathe- matical representation of physical phenomena, or it may arise as follows: Suppose a cus- tomer’s service time has an exponential distribution with parameter 8. The random vari- able T, the total time to service n (independent) customers, has a gamma distribution with parameters n and 8 (replacing a and {3, respectively); that is,

INTRODUCTION TO OPERATIONS RESEARCH-0564

Note that when n = 1 (or a = 1) the gamma density becomes the density function of an exponential random variable. Thus, sums of independent, exponentially distributed ran- dom variables have a gamma distribution.

Another important distribution, the chi square, is related to the gamma distribution. If X is a random variable having a gamma distribution with parameters {3 = 1 and a = v/2 (v is a positive integer), then a new random variable Z = 2X is said to have a chi- square distribution with v degrees of freedom. The expression for the density function is given in Table 24.1 at the end of Sec. 24.8.

The Beta Distribution

A continuous random variable whose density function is given by

INTRODUCTION TO OPERATIONS RESEARCH-0565

is known as a beta-distributed random variable. This density is a function of the two pa- rameters a and {3, both of which are positive constants. A graph of a typical beta density function is given in Fig. 24.12.

Beta distributions form a useful class of distributions when a random variable is re- stricted to the unit interval. In particular, when a = {3 = 1, the beta distribution is called the uniform distribution over the unit interval. Its density function is shown in Fig. 24.13, and it can be interpreted as having all the values between zero and 1 equally likely to oc- cur. The CDF for this random variable is given by

INTRODUCTION TO OPERATIONS RESEARCH-0566

INTRODUCTION TO OPERATIONS RESEARCH-0567INTRODUCTION TO OPERATIONS RESEARCH-0568

In this case -(b - JL)/u is positive, and FX(b) = a is thereby read from the table by entering it with -(b - JL)/u. Thus, suppose it is desired to evaluate the expression

P{2 < X ::: 18} = FX(18) - FX(2).

FX(18) has already been shown to be equal to 1 - 0.0228 = 0.9772. To find FX(2) it is first noted that (2 - 10)/4 = -2 is negative. Hence, Table A5.1 is entered with Ka = +2, and FX(2) = 0.0228 is obtained. Thus, FX(18) - FX(2) = 0.9772 - 0.0228 = 0.9544.

As indicated previously, the normal distribution is a very important one. In particular, it can be shown that if X1, X2, . . . , Xn are independent,1 normally distributed random 1The concept of independent random variables is introduced in Sec. 24.12. For now, random variables can be considered independent if their outcomes do not affect the outcomes of the other random variables.

INTRODUCTION TO OPERATIONS RESEARCH-0569

Comments

Post a Comment

Popular posts from this blog

DUALITY THEORY:THE ESSENCE OF DUALITY THEORY

By -
Image
THE ESSENCE OF DUALITY THEORY Given our standard form for the primal problem at the left (perhaps after conversion from another form), its dual problem has the form shown to the right. Thus, with the primal problem in maximizatio n form, the dual problem is in minimization form instead. Furthermore, the dual problem uses exactly the same pa r amete r s as the primal problem, but in different locations, as summarized below. 1. The coefficients in the objective function of the primal problem are the right-hand sides of the functional constraints in the dual problem. 2. The right-hand sides of the functional constraints in the primal problem are the coef- ficients in the objective function of the dual problem. 3. The coefficients of a variable in the functional constraints of the primal problem are the coefficients in a functional constraint of the dual problem. To highlight the comparison, now look at these same two problems in matrix notation (as introduced at the beginn...
Read more

NETWORK OPTIMIZATION MODELS:THE MINIMUM SPANNING TREE PROBLEM

By -
Image
THE MINIMUM SPANNING TREE PROBLEM The minimum spanning tree problem bears some similarities to the main version of the shortest-path problem presented in the preceding section. In both cases, an undi r ecte d and connected network is being considered, where the given information includes some mea- sure of the positive length (distance, cost, time, etc.) associated with each link. Both problems also involve choosing a set of links that have the shortest total length among all sets of links that satisfy a certain property. For the shortest-path problem, this property is that the chosen links must provide a path between the origin and the destination. For the minimum spanning tree problem, the required property is that the chosen links must provide a path between ea c h pair of nodes. The minimum spanning tree problem can be summarized as follows: 1. You are given the nodes of a network but not the links. Instead, you are given the po- tential links and the positive length for each ...
Read more

INTEGER PROGRAMMING:THE BRANCH-AND-CUT APPROACH TO SOLVING BIP PROBLEMS

By -
TH E BRANCH-AND-CUT APPROACH TO SOLVING BIP PROBLEMS Integer programming has been an especially exciting area of OR since the mid-1980s be- cause of the dramatic progress being made in its solution methodology. Background To place this progress into perspective, consider the historical background. One big break- through had come in the 1960s and early 1970s with the development and refinement of the branch-and-bound approach. But then the state of the art seemed to hit a plateau. Rel- atively small problems (well under 100 variables) could be solved very efficiently, but even a modest increase in problem size might cause an explosion in computation time be- yond feasible limits. Little progress was being made in overcoming this exponential growth in computation time as the problem size was increased. Many important problems arising in practice could not be solved. Then came the next breakthrough in the mid-1980s, with the introduction of the b r an c h- and-cu t approach to sol...
Read more