ࡱ> dfcu bjbj ;hhF:tt8l0i"F$F$F$0000000$1m4<0QF$$@F$F$F$<0{ 0...F$H0.F$0...izE%.0000.5~*5.5.F$F$.F$F$F$F$F$<0<0,F$F$F$0F$F$F$F$5F$F$F$F$F$F$F$F$F$t }: Lecture 17 STAT 305 [NOTE: This lecture is Section 7 of my ongoing manuscript] 7. FUNCTIONS OF RANDOM VARIABLES We begin this topic by reminding the reader, once again, of the nature of a random variable. Definition 7.1. An n-dimensional (n-D) variable,  EMBED Equation.3 is a composite action, composed of n individual actions, that, when performed, yields a vector measurement value  EMBED Equation.3 . The set of all measurable values that X can take on is called the sample space for X. If the sample space includes more than a single element, then X is said to be an n-D random variable. It must be stressed that X is an action, and that x is a vector of numbers. If this distinction is kept in mind, then whenever numbers are presented, one is more likely to ask the question: what actions resulted in these numbers? As the title of this discussion notes, we are here concerned with functions of random variables; which might pose the following question: What is a function? In fact, a function is also an action. The terms action, function, algorithm, etc. are fairly synonymous. They are operations that, when performed, yield viewable results. With this in mind, we state what some might consider to be obvious: Because a function of a random variable is an action, and the random variable itself is an action, a function of a random variable, being a composite action, is, itself, a random variable. Numerous examples of functions of Bernoulli random variables were given in the notes on n-D Bernoulli random variables. In this set of notes, we will look at functions of other random variables. In an effort to tie this topic to the topic of distribution models for random variables, we will begin by looking at how one might simulate them. This will entail a rather strange looking function of a random variable. 7.1 Simulation of Measurements of a Random Variable WHY? Because when you design strategies based on measurements alone, if a strategy does not perform well it is not always clear as to the reason(s) for poor performance. Example 7.1. You want to design a course performance evaluation procedure. If you design it based on exam and homework performance data for a limited sample size, the statistics used in the procedure (e.g. mean, variance, PDF) may entail notable uncertainty. This could result in a procedure that is not well-suited to the next class. Or, it could be that the procedure itself is flawed. Simulation of 1-D Random Variables using a U(0,1) Random Number Generator Case1: X is a Discrete Random Variable with a finite Sample Space Recall how the uniform random number generator, U, was used to simulate a 2-D Bernoulli random variable,  EMBED Equation.3 . Since the sample space includes only four elements, with probabilities  EMBED Equation.3 , having specified these probabilities, it was straightforward to simulate X. While it may not have seemed so, we had, in fact, defined X to be a function of U. Specifically, we defined it via equivalence of the following events:  EMBED Equation.3  This same procedure can be used to simulate any discrete random variable. Consider the following example. Example 7.2 Suppose that you want to simulate the grades of a class of 20 students, where the possible grades are A=4, B=3, C=2, D=1, and F=0. Suppose further, that you specify the following probabilities associated with this X:  EMBED Equation.3  The five values that X can take on will be dependent on the values that U~Uniform[0,1] takes on. The equivalent events can be chosen to be  EMBED Equation.3  The Matlab code for a simulation of X is given in the Appendix. Below are 3 simulations obtained from that code. >> ex71 c = 0 2 10 5 3 >> ex71 c = 1 1 11 6 1 >> ex71 c = 0 1 8 7 4 What is the value of such simulations? By showing someone who is not familiar with probability and statistics the above results, they can more easily see the possibilities, either over multiple semesters of the same course, or of three courses with the same probability model in a given semester. This type of information can be useful in budget planning (e.g. how many students might retake the course), and in predicting program success outcomes. Case 2: X is a Continuous Random Variable We will begin this section with a function of a random variable that is useful in simulating measurements. Using the Uniform Random Number Generator to Simulate Measurements of a Continuous Random Variable Recall that the cdf of a U(0,1) random variable, U, is  EMBED Equation.3  (1) Notice that the U(0,1) random variable cdf has the unique defining property (1); that is, the quantity on the right side of the inequality is exactly the same quantity that the probability of that event is equal to. Now let X be a random variable with CDF  EMBED Equation.3 . Since X is assumed to be purely continuous,  EMBED Equation.3 is strictly monotonically increasing. Thus,  EMBED Equation.3  (2) QUESTION: What do equations (1) and (2) have in common? ANSWER: Both are the cdf of a U(0,1) random variable. In (1) it is U, and in (2) it s  EMBED Equation.3 . Remark. The right equality in (2) merits some discussion. First, the fact that  EMBED Equation.3 is a monotonic function means that the inverse function  EMBED Equation.3 exists. In simple terms, it means that one can go back and for the between x and  EMBED Equation.3  without confusion. This is illustrated below.   EMBED Equation.3   EMBED Equation.3  x1 x2 Figure 17.1 Because  EMBED Equation.3 is a monotonic continuous function, one can as easily determine x2 given  EMBED Equation.3 as one can determine  EMBED Equation.3  given x1. Notice that by performing exactly the same operation  EMBED Equation.3 to both sides of the inequality  EMBED Equation.3 the actual set  EMBED Equation.3 remains the same; that is, it is no different that  EMBED Equation.3 , which is, in my opinion, a strange looking event. Specifically, the quantity  EMBED Equation.3 is not a cdf; rather, it is a random variable. If we define the random variable  EMBED Equation.3 , and  EMBED Equation.3 , then (2) is identical to (1). Hence, we can conclude that for any given CDF  EMBED Equation.3 , the uniform random number generator generates samples of  EMBED Equation.3 . Suppose that the number u has been generated. Since  EMBED Equation.3 is strictly monotonic, its inverse function exists. Hence, (2) becomes  EMBED Equation.3  (3) From (3) we see that each number u corresponds to a unique number x, such that the probability relation (3) is maintained. Example 7.2 Use (3) to simulate 1000 measurements of  EMBED Equation.3 . The pdf for X is  EMBED Equation.3 . (4a) It follows from calculus that the cdf is:  EMBED Equation.3 . (4b) To find function, or algorithm, EMBED Equation.3 amounts to solving (4b) for x in terms of y. To this end:  EMBED Equation.3 . (5) A scaled histogram of the simulation, overlaid against the model pdf is given in Figure 2. The Matlab code is given in the Appendix.  Figure 2. Comparison of the scaled histogram and the theoretical pdf for  EMBED Equation.3  % Example 17.3 Very often in engineering, it is the square of a measurement that is of interest. For example, in studying the power consumption associated with a solar panel, it is the square of voltage that is pertinent. In the case of wind resistance experienced by an automobile traveling against the wind at ~50 mph it is the square of the relative wind speed that represents the force on the vehicle. [Although, at higher speeds this square law becomes a fourth power law.] Hence, in this example our interest is in the cdf and pdf of the random variable Y = X2. To begin, lets assume that the sample space for X is the entire real line:  EMBED Equation.3 . Furthermore, we will assume that X is a purely continuous random variable. We will begin by addressing cumulative events, since these are directly related to cumulative probability. It should be clear that the following two events are one and the same:  EMBED Equation.3 . (6a) We may also write this as  EMBED Equation.3 . (6b) An example of the two sets for y=4 is shown below.  0  EMBED Equation.3  4 -2  EMBED Equation.3  2 Again, while the first expression is more compact, the second does not require that the reader know the difference between an upper case quantity (i.e. a random variable), and a lower case quantity (i.e. a number). It follows immediately that  EMBED Equation.3 . (7) The leftmost equality in (7) is simply the definition of a cdf. The rightmost equality is due to the simple fact that the probability contained in the interval  EMBED Equation.3  is simply all the probability to the left of  EMBED Equation.3  minus all the probability to the left of  EMBED Equation.3  . [Note that, because we have assumed that X is a purely continuous random variable, it makes no difference if we replace any of the above inequalities by strict inequalities; that is to say,  EMBED Equation.3 .] We have essentially completed the conceptual part of the problem. For any chosen cdf for X, we see from (7) that the cdf for Y is given as the difference between two cdf values for X. What remains is mathematics. It can be as simple as a picture, or as complicated as calculus, depending on our assumed form of the cdf for X. We now demonstrate each of these cases. Case 1: Assume that X ~ Uniform[-100 , 100]. The pdf for X is  EMBED Equation.3 and so the cdf for X is  EMBED Equation.3 . The sample space for Y = X2 is SY = [0 , 104]. From (7) we have  EMBED Equation.3 . (8a) The pdf for Y is the derivative of (8a). Specifically, we find that  EMBED Equation.3 . (8b) Plots of (8) are given below. Notice that since  EMBED Equation.3 , the pdf plot was necessarily truncated.  Figure 3. Plots of the cdf (top) and the pdf (bottom) corresponding to (8a) and (8b), respectively. Remarks. It is worth noting that the pdf of Y approaches infinity as y approaches zero. Even so, it is a valid pdf. The behavior of the pdf for larger values of y is consistent with the behavior of the cdf. Specifically, notice that the cdf is accumulating probability at a relatively constant rate for larger values of y. The slope is approximately (1 0.7) / (10000-5000), or ~6(10-5). This is consistent with what we see in a zoomed version of the pdf:  Figure 4. A zoomed region of the pdf for Y, in comparison to the cdf-estimated slope of ~6(10-5). The point here is to emphasize the fact that the pdf is a rate function: it is a measure of the rate at which probability accumulates as x increases. One final remark concerning this case is in order. The fact that calculus was used to obtain the pdf expression (8b) should not cause those with no calculus background any great concern. As was shown above, an estimate of the pdf in the interval (5000, 10000) was obtained using only the definition of a slope (i.e. the rise over the run). The cdf was obtained directly from the knowledge of events. No calculus (or even algebra) was used. From the cdf one could proceed to get a good idea of the shape of the pdf by simply measuring the slope at a variety of x-values, and then drawing a smooth curve through those points. To simulate Y we can compute the inverse function associated with (8a):  EMBED Equation.3 . And so, to simulate a measurement of Y, we first simulate a measurement of U ~ Uniform(0,1), then we multiply that measurement by 100 and square the result. Alternatively, we could begin by simply simulating a measurement of X. Since X ~ Uniform(-100, 100), we can express it as  EMBED Equation.3 . Multiplying U by 200 changes the sample space from [0, 1] to [0, 200]. Subtracting 100 shifts this sample space to [-100, 100]. So, to simulate a measurement of Y, we would first simulate a measurement of U. Next, we would transform it to a simulation of EMBED Equation.3 . And finally, we would obtain Y = X2. The histogram-based (using 10,000 simulated measurements) and the true, 8(b), pdfs are compared in Figure 5 below. This figure validates the pdf expression 8(b). Note that since 8(b) was computed only at the bin centers (and then interpolated), it was not computed for values of y < 500.  Figure 5. histogram-based (using 10,000 simulated measurements)(blue) and the true, 8(b), (black) pdfs. Case 2: Assume that X is Normal( 0,  EMBED Equation.3 ) . Then, the cdf for Y = X2 is still given by (7), but where now  EMBED Equation.3 . (9) In contrast to (8a), there is no closed form for (9). It is for that reason that almost every textbook on the subject provides a table of cdf values for the N( 0, 1) random variable. Such a table could be used, along with (7), to obtain the value of the cdf for Y at a variety of y-values, and then connect the points with a smooth curve. Having the cdf, the pdf could then be approximated in a manner exactly as was described in Case 1. We will now obtain the form of the pdf for Y using calculus. To begin, we will differentiate (7):  EMBED Equation.3 . (10) Now, using the chain rule for differentiation, we have  EMBED Equation.3  , where we have defined EMBED Equation.3 . Notice that  EMBED Equation.3 . Hence, we have  EMBED Equation.3 . Similarly,  EMBED Equation.3 . Hence, (10) becomes  EMBED Equation.3 . (11) Special Case: Suppose that we now define the random variable  EMBED Equation.3 . Then the cdf of W is:  EMBED Equation.3 . Just as we did above, we can apply the chain rule to obtain the pdf for W:  EMBED Equation.3 . (12) Notice that  EMBED Equation.3 , where Z ~ Normal(0 , 1). Hence, (12) is the pdf of the square of the standard normal random variable, Z. The random variable W = Z2 has a name. It is called the chi-squared random variable with one degree of freedom. We will complete this example with a comparison of the pdf for Y = X2 under the assumption that X has a Uniform[-10,10] pdf versus that X has a zero mean normal pdf with  EMBED Equation.3 . The reason for this comparison is that very often it is assumed that random variables have a normal pdf. However, if one has no information about X other than it has an essential range of [-10, 10], then the uniform pdf is a more logical choice. The two pdfs are shown in Figure 6.  Figure 6. Comparison of pdfs for Y = X2, assuming X ~Uniform[-10,10] (blue) versus assuming X ~Normal(0,3.332). The top plot includes the entire pdfs, and the lower two are zoomed regions. We see that the differences between the pdfs are dramatic. Under the normal assumption the majority of probability is concentrated below y=25. In contrast, under the uniform assumption, probability is much more evenly distributed over the entire region [0, 10,000]. In statistical jargon, one would observe that the tail associated with the uniform pdf assumption is much fatter than that resulting from the normal assumption. As we will see, a fat tail can result in a much larger than expected mean value. The difference in model tail thicknesses, itself, can be crucial in obtaining reliable probabilities of extreme events. For example, from lowest plot in Figure 6, we have two model-based estimates of Pr[9000 < X < 90100]: The uniform-based estimate is ~ (5x10-6)(10) = 5x10-5. Since the normal-based pdf value is ~ 1/10th that of the uniform in this region, the normal based probability is ~ 5x10-6; that is, these two models differ by an order of magnitude in their prediction of the probability of a tail event. % APPENDIX % PROGRAM NAME: ex71.m p=[0.05 0.10 0.35 0.3 0.2]; % Grade probabilities %================================== F=[p(1) zeros(1,4)]; % CDF for X for k=2:5 F(k)=F(k-1)+p(k); end % ================================ N=20; % Number of students taking the test x=zeros(1,N); c=zeros(1,5); u=rand(1,N); for n=1:N if u(n) < F(1) x(n)=0; c(1)=c(1)+1; elseif u(n) < F(2) x(n)=1; c(2)=c(2)+1; elseif u(n) < F(3) x(n)=2; c(3)=c(3)+1; elseif u(n) < F(4) x(n)=3; c(4)=c(4)+1; else x(n)=4; c(5)=c(5)+1; end end % PROGRAM NAME: expsim.m theta=0.2 % COMPUTE THE Exp PDF OVER THE INTERVAL [0,10] x=0.1:0.01:1; fe=(1/theta)*exp(-theta^-1*x); % ======================================== % SIMULATE 1000 MEASUREMENTS OF X u=rand(1,1000); xsim=-theta*log(1-u); % COMPUTE SCALED HISTOGRAM bw=0.1; % Bin width bvec=bw/2:bw:1-bw/2; N=hist(xsim,bvec); fh=(bw*1000)^-1 *N; % ======================================== figure(1) bar(bvec,fh) hold on plot(x,fe,'bl') xlabel('x') ylabel('PDF Estimate') Ucu      % ' 4 5 H I J K n o  # ( * i j z жжЧЕжЧraжжжж!jhhCh-CJ EHUaJ #j5I hhCh-CJ UVaJ !jhhCh-CJ EHUaJ #j4I hhCh-CJ UVaJ jhhCh-CJ UaJ hhCh-6CJ aJ hhCh-56CJ aJ hhCh-CJ aJ hhChd75CJ aJ hhCh-5CJ aJ hhCh?5CJ aJ %TU    cdFGZ["#;gd&^gd-z | }  P w      cdhжЈznbVJhhCh5CJ aJ hhCh&^5CJ aJ hhChd75CJ aJ hhChC6CJ aJ hhChC56CJ aJ hhChCCJ aJ hhChr6CJ aJ hhChrCJ aJ hhChon6CJ aJ hhCh-5CJ aJ hhCh-56CJ aJ hhCh-CJ aJ hhChon5CJ aJ hhChon56CJ aJ hhChonCJ aJ EFGNRT&Y[󟐻znzz_zM#j>I hhChd CJ UVaJ jhhChd CJ UaJ hhChd 6CJ aJ hhChd CJ aJ hhChCCJ aJ hhCh&^56>*CJ aJ hhCh&^56CJ aJ hhCh&^5>*CJ aJ hhCh&^CJ aJ hhChd75CJ aJ hhCh?5CJ aJ hhChC5CJ aJ hhC5CJ aJ hhCh&^5CJ aJ $%89:;!"#$789:;<hkɷɚɚɚɈwlaUahhChd76CJ aJ hhChd7CJ aJ hhCh&^CJ aJ !j hhChd7CJ EHUaJ #j>I hhChd7CJ UVaJ hhChd 6CJ aJ !jghhChd CJ EHUaJ #j=I hhChd CJ UVaJ hhChd CJ aJ hhChCCJ aJ jhhChd CJ UaJ !jhhChd CJ EHUaJ ;<de67?aiz{gd gd&^!#/cd56軪蟐~meZNZNZhhCh 6CJ aJ hhCh CJ aJ hhCCJ aJ !jghhChhCCJ EHUaJ #jL hhChhCCJ UVaJ jhhCh CJ UaJ hhCh CJ aJ !j hhChd7CJ EHUaJ #jBI hhChd7CJ UVaJ jhhChd7CJ UaJ hhChd76CJ aJ hhChd7CJ aJ hhChd75CJ aJ 6y{vƸܒ~r~~j~[~I#jL hhCh 4 CJ UVaJ jhhCh'dCJ UaJ h 4 CJ aJ hhCh'd6CJ aJ hhCh'dCJ aJ h 4 6CJ aJ hhCh&^CJ aJ hhCh&^56>*CJ aJ hhCh&^5CJ aJ hhCh&^56CJ aJ hhC56CJ aJ hhChCJ aJ hhCh56CJ aJ hhChd7CJ aJ hhCh CJ aJ uv    0*gd 4  0*gd'd /gd 4 gd&^ 145=MUV   !"*+OPQde̷̬ԠԎ}ԠrcrQ#jZG hhChc`CJ UVaJ jhhChc`CJ UaJ hhChc`CJ aJ !jhhCh 4 CJ EHUaJ #jZG hhCh'dCJ UVaJ hhCh'd6CJ aJ h5h'dCJ aJ h56CJ aJ h5h56CJ aJ h5CJ aJ hhCh'dCJ aJ jhhCh'dCJ UaJ !j4hhCh 4 CJ EHUaJ efgj .1\]ԽԡԒoԵg\SgSgJgJh56CJ aJ h55CJ aJ hhCh5CJ aJ h5CJ aJ !jhhCh 4 CJ EHUaJ #j͇L hhCh 4 CJ UVaJ jhhCh'dCJ UaJ h 4 CJ aJ hhCh 4 6CJ aJ h'dCJ aJ hhCh'dCJ aJ hhChc`6CJ aJ hhChc`CJ aJ jhhChc`CJ UaJ !jhhCh 4 CJ EHUaJ ]op'(;<̻鳨tbQt?#j݈L hhCh 4 CJ UVaJ !j hhCh 4 CJ EHUaJ #jZG hhCh 4 CJ UVaJ hhCh 4 CJ aJ jhhCh 4 CJ UaJ h 4 CJ aJ h 4 5CJ aJ h55CJ aJ h5h5CJ aJ h*CJ aJ !jhhCh*CJ EHUaJ #jL hhCh*CJ UVaJ hhCh5CJ aJ jhhCh5CJ UaJ h5CJ aJ <=>    ñ׎|kYH!j(h 4 h#CJ EHUaJ #jL hhCh#CJ UVaJ !j&h 4 h#CJ EHUaJ #jL hhCh#CJ UVaJ "jh4CJ UaJ mHnHu!j$hhCh 4 CJ EHUaJ #jZG hhCh 4 CJ UVaJ hhCh 4 CJ aJ h 4 6CJ aJ h 4 CJ aJ jhhCh 4 CJ UaJ !j/"h 4 h 4 CJ EHUaJ d!e!##8#9###$$$h$i$$$ /gd* l0*gd#  gd 4  0*gd'd$-.ABCDźźtźbQ!j/h#h4CJ EHUaJ #j+hhCh#CJ EHUaJ #jZG hhCh#CJ UVaJ hhCh#CJ aJ jhhCh#CJ UaJ h#5CJ aJ h45CJ aJ h#CJ H*aJ h#6CJ aJ h#CJ aJ   % & ' ( H I \ ] ^ _ n o °ܓ܄sܓdSܓDjL h4CJ UVaJ !j_6h4h4CJ EHUaJ jL h4CJ UVaJ !j44h4h4CJ EHUaJ jyL h4CJ UVaJ jh4CJ UaJ !j1hhCh4CJ EHUaJ #j!C!S!b!c!d!e!!!!!ø҉ҀҀumbSbA#j]G hhChc`CJ UVaJ jhhChc`CJ UaJ hhChc`CJ aJ h 4 CJ aJ h5h4CJ aJ h56CJ aJ h556>*CJ aJ !j;hhCh5CJ EHUaJ #jL hhCh5CJ UVaJ hhCh5CJ aJ jhhCh5CJ UaJ h5CJ aJ h4CJ aJ jh4CJ UaJ !j8h4h4CJ EHUaJ !!!!!!!!!" """""Z"["n"o"p"q"""""""±ԟ|k_TETjhhChC CJ UaJ hhChC CJ aJ hhChc`6CJ aJ !jDhhCh*CJ EHUaJ #j^G hhChc`CJ UVaJ !jBhhCh*CJ EHUaJ #jZG hhChc`CJ UVaJ !j0@hhCh*CJ EHUaJ #j]G hhChc`CJ UVaJ hhChc`CJ aJ jhhChc`CJ UaJ !j=hhCh*CJ EHUaJ """"####0#1#2#3#Z#[#{#|########·¥ˆˆ}rfrWLhhChqCJ aJ jhhChqCJ UaJ hhChB5CJ aJ hhChBCJ aJ hhCh'dCJ aJ hhChC 6CJ aJ !jHhhCh*CJ EHUaJ #j_G hhChC CJ UVaJ hhChc`CJ aJ hhChC CJ aJ jhhChC CJ UaJ !jFhhCh*CJ EHUaJ #jZG hhChC CJ UVaJ #$$$$$ $ $$$J$K$^$_$`$a$i$$$$$$$$$$%%*%·«Ÿ|t«bQ!jQhhCh*CJ EHUaJ #jJI hhChqCJ UVaJ h*CJ aJ !jNhhCh*CJ EHUaJ #jBI hhChqCJ UVaJ hhChq6CJ aJ h*hq6CJ aJ hhChBCJ aJ hhChqCJ aJ jhhChqCJ UaJ !j5LhhCh*CJ EHUaJ #jI hhChqCJ UVaJ $$$$g%h%i%%%$&&&*&&&&)++i++++++ /7$8$H$gd83 7$8$H$gd  $ 0*a$gdd /gd* 0*gd'd*%+%,%-%I%J%W%X%i%n%o%p%%%%%%%%%&&(&*&>&&¶¶®œ̀€teYNhhChdCJ aJ hhChd5CJ aJ jYhhChdCJ UaJ hhChZr6CJ aJ hhChZrCJ aJ !jVhhCh*CJ EHUaJ #jI hhChZrCJ UVaJ h*CJ aJ hhChq6CJ aJ hhChqCJ aJ jhhChqCJ UaJ !jsThhCh*CJ EHUaJ #jI hhChqCJ UVaJ &&&&&&&&& ''X'f'/)9)D)U)))))))))))))*𺯣xxxxxxoxg[gjh7CJ UaJ h7CJ aJ h]CJ H*aJ h]6CJ aJ h]CJ aJ h]5B* CJ aJ ph""h]h]5CJ aJ h]B* CJ aJ ph""hhChqCJ aJ hcCJ aJ !jdhhChcCJ EHUaJ #jI hhChdCJ UVaJ hhChdCJ aJ jhhChdCJ UaJ *****6*7*8*E*a*d*K+L+_+`+a+b+c+h++++++++++++++÷î˔{{l[{{î!jIlh7h83CJ EHUaJ j.L h83CJ UVaJ h83CJ aJ !jqih7h7CJ EHUaJ j-L h7CJ UVaJ hB|h7CJ aJ hB|6CJ aJ hB|hB|6CJ aJ hB|CJ aJ h7CJ aJ jh7CJ UaJ !j.gh7h7CJ EHUaJ jb,L h7CJ UVaJ ++,,,,q-r-----///@1A1 2 20212u2v22 /7$8$H$gdV 7$8$H$gd;8t /7$8$H$gd83 7$8$H$gd +++,,,?,@,A,B,J,K,j,l,n,u,v,w,,,,,,,--------W.X.k.˺棒ڲ~o^~U~h836CJ aJ !jhth83h83CJ EHUaJ jf0L h83CJ UVaJ jh83CJ UaJ h83CJ aJ !jqh7hB|CJ EHUaJ j3L hB|CJ UVaJ h;8tCJ aJ !jvohB|hB|CJ EHUaJ j3L hB|CJ UVaJ jhB|CJ UaJ hB|CJ aJ "jh;8tCJ UaJ mHnHu k.l.m.n............. /!//////////#0˿˰˿ːvn_NFFh;8tCJ aJ !j_hB|hB|CJ EHUaJ j2L hB|CJ UVaJ h83CJ aJ hB|6CJ aJ !j|hB|hB|CJ EHUaJ j2L hB|CJ UVaJ !jzhB|hB|CJ EHUaJ j1L hB|CJ UVaJ jhB|CJ UaJ hB|CJ aJ jh83CJ UaJ !jxh83hB|CJ EHUaJ j1L hB|CJ UVaJ #0&0+0,0G0J0O0P0x0{000 1111A1F1I1U1`1d1e1j1k1n1r1u1v1z1|111111111111111ƽҽұҢҽҽұ҂j9L hVCJ UVaJ !jρh%hVCJ EHUaJ j8L hVCJ UVaJ jh%CJ UaJ h%6CJ aJ h%h%6CJ aJ h%CJ aJ h;8t5CJ aJ h;8t56CJ aJ h;8tCJ aJ h;8t6CJ aJ ,1111111111111112 222&2'2(2)25282=2>2v22222ѵڭڡڒxxpaj:L h CJ UVaJ h CJ aJ hV6CJ aJ !jȆh83hVCJ EHUaJ j9L hVCJ UVaJ jhVCJ UaJ h;8tCJ aJ h%6CJ H*aJ h%CJ H*aJ h%CJ aJ h%6CJ aJ hVCJ aJ jh%CJ UaJ !j?h%hVCJ EHUaJ 22222222222223"3#3$3%3&3/3=3@3O3R3333333333344,4-4U4Y4x4{4444ʾʯʕʊ{rʕʕrʕʕʕʕʕʕʕʕʕhF5CJ aJ jhFhXCCJ UaJ hFh CJ aJ hF6CJ aJ !jhFhFCJ EHUaJ jU>L hFCJ UVaJ jhFCJ UaJ hFCJ aJ hVCJ aJ h CJ aJ jhVCJ UaJ !jhVh CJ EHUaJ ,222#3$3&333T5V555O6P688 9 9E9F9: /7$8$H$gdXC$ /7$8$H$a$gdXC /7$8$H$gdF$ /7$8$H$a$gdF /7$8$H$gdV445 5 5O5R5S5T5U5V5`5w5z55555555566N6O6P6X6u66627577788N8Q88888vvvvvvnhXC9CJ aJ h] 6CJ aJ hCJ aJ h] CJ aJ h3#hXC56CJ aJ h3#h3#5CJ aJ h3#h3#56>*CJ aJ h3#h3#56CJ aJ hXC5CJ aJ j hXChXCCJ UaJ hXC6CJ aJ hXCCJ H*aJ hFCJ aJ hXCCJ aJ *88 9,9-9@9A9B9C9D9E9k9n999':(:0:4:]:^:q:r:s:t::::;;C;D;t;u;;˺斅}t}t}h}jh bCJ UaJ h b6CJ aJ h bCJ aJ !jyhp.hp.CJ EHUaJ jL hp.CJ UVaJ jhp.CJ UaJ hp.6CJ aJ !jh83hp.CJ EHUaJ jӆL hp.CJ UVaJ jhXC9CJ UaJ hp.CJ aJ hXC9CJ aJ hXC96CJ aJ ";;;;;;;;<<%<&<A<D<<<<<<<<<<9=>=?=@=A=B=G=J=V=¹˱˱}rg^VhCJ aJ h5CJ aJ h56CJ aJ hVh CJ aJ hXC9CJ aJ h\25CJ aJ jh\2h\2CJ UaJ h\2h\2CJ aJ h\26CJ aJ h\2CJ aJ h bCJ H*aJ h b6CJ aJ h bCJ aJ jh bCJ UaJ !jƻhp.h bCJ EHUaJ jL h bCJ UVaJ :&<'<<<@=A=B=====??@@T@U@@@ /7$8$H$gd 7$8$H$gd  /7$8$H$gdV /7$8$H$gd\2$ /7$8$H$a$gd\2 /7$8$H$gdXCV=W=[=a=f=g=z={=|=}================J>>>>>>>>???W?Z?`?c?????????xxxxxxxxxxhq6CJ aJ hqCJ aJ !jhha fCJ EHUaJ jUL ha fCJ UVaJ hACJ aJ h83CJ aJ hCJ H*aJ !jThhCJ EHUaJ jOL hCJ UVaJ jhCJ UaJ hCJ aJ h6CJ aJ /?6@7@J@K@L@M@S@T@@@@@@@@@@@@@@@@AA$A%A}n]NjRUL hACJ UVaJ !jhAhACJ EHUaJ jTL hACJ UVaJ hiqCJ aJ !jhAhACJ EHUaJ jSL hACJ UVaJ jhACJ UaJ hCJ aJ hACJ aJ hqhqCJ aJ !jIhqhACJ EHUaJ j2SL hACJ UVaJ jhqCJ UaJ hqCJ aJ @@@)A*A:A;AcAnAoAAAAAAcBdBBBBBBC CDD /7$8$H$gdW /7$8$H$gdiq 7$8$H$gd %A&A'A)A;AJAKA^A_A`AaAbAcA}A~AAAAAAAAAAAAAAAʾүʒʃrʒcRJhXOCJ aJ !jh,hiqCJ EHUaJ jL hiqCJ UVaJ !j^h,hiqCJ EHUaJ jSL hiqCJ UVaJ jhiqCJ UaJ !jh,hiqCJ EHUaJ j\L hiqCJ UVaJ jha fCJ UaJ hiqCJ aJ ha fCJ aJ hACJ aJ jhACJ UaJ !jlhAhACJ EHUaJ ABB3B4B5BHBIBJBKBMBVBYB]B^BbBcB~BBBBBBBBBBBBCCCַ֜p_ߜ!jhWhWCJ EHUaJ jL hWCJ UVaJ jhWCJ UaJ hiqCJ aJ hW6CJ aJ hWCJ aJ $jhWhW5CJ EHUaJ jہL hW5CJ UVaJ jhW5CJ UaJ hW5CJ aJ hWhWCJ aJ hiq5CJ aJ hiq56>*CJ aJ CCCCCC C,C-C@CACBCCCECKCUCqCtCCCCCCCDDDVDYD˻wwwwlaYQHh<6CJ aJ hXOCJ aJ h7CJ aJ hXC9hWCJ aJ hXC96CJ H*aJ hXC96CJ aJ $jhXC9hXC95CJ EHUaJ jL hXC95CJ UVaJ hXC95CJ aJ jhXC95CJ UaJ hXC9CJ aJ hACJ aJ hWCJ aJ jhWCJ UaJ !jhWhXC9CJ EHUaJ jL hXC9CJ UVaJ YD^DcDdDDDDDDDDDDDDDDDGEJEuEvEEEEEEEEFF FFF#F(F)F4F>F^FgFnFoFqF˺}}t}}th]CJ H*aJ h]6CJ aJ h]CJ aJ h]5CJ aJ h]B* CJ aJ ph""j hEhECJ UaJ h7CJ aJ !jh<h{CJ EHUaJ j`L h{CJ UVaJ jh<CJ UaJ h<CJ H*aJ h<6CJ aJ h<CJ aJ +DEFFFFHHKKKKL;L\LfL|LLLLLLLMM 7$8$H$gd] $7$8$H$a$gd] $7$8$H$a$gdE 7$8$H$gd qFFFFFFFKGMGOGH"HHIIIIIIIIIIJJJJKKKKKK˺|pdYhhCh CJ aJ hhCh|5CJ aJ hhCh~l5CJ aJ h *ThwVB* CJ aJ ph""h *Th *TCJ H*aJ h *T6CJ aJ h *TCJ H*aJ h *TCJ aJ hD6CJ aJ hDCJ aJ h~l6CJ aJ h~lCJ aJ h]6CJ aJ h]CJ aJ hEh]CJ aJ hE6CJ aJ hECJ aJ KK N!N#N9NOðİưǰɰʰ̰ͰӰ԰հְװڰ۰Ҽh *TmHnHujhhCUhhCh\hd hd0Jjhd0JUh,Kjh,KUUh~lhd5CJ aJ hhChdCJ aJ hhChY$pCJ aJ hhCh CJ aJ h~lh 5CJ aJ !M$M9MPM`MuMMMMMMMMNNN N!N:NDNsNNNNNNO0O 7$8$H$gdd 7$8$H$gd 0ODOYOlOOOOOOOOO9?°ðŰưȰɰ˰̰հ &`#$gd\ 7$8$H$gddtitle('Comparison of Scaled Histogram and Exp(0.2)PDFs') pause % Since the fit is not very good, shrink a to make it less exponential a1=1.3*a; b1=.7*b; fg1=gammapdf(x,a1,b1); plot(x,fg1,'r')     PAGE  17. PAGE \* MERGEFORMAT 17 հְװ 7$8$H$gdd$a$h]hgdd9&P10:phC= /!"#$% Dd hb  c $A? ?3"`?2UFJDV LD`!UFJDV Lr @|vxcdd`` @c112BYL%bpu?}ר?VDAa dB?Q >dBܤD._m+apANNLLJ% {@2u(2tA4T}b@#3Xm^PDd hb  c $A? ?3"`?2N$tn.+v`!nN$tn.+ @8>|<xcdd``ved``baV d,FYzP1n:&&n! KA?H1Z ㆪaM,,He`H @201d++&1X +|-“NHq50qV/A|8? /WWBܣ<02pw=`L! 7X0AFz0.w0gȄ_ ?8f3v0o8-021)W2lC D,ĀF`^vDd b  c $A? ?3"`?2,C;rC+ `!C;rC+:!@ 3PVxڥk@d&6.6AZjuŋR/TV)h-lkm]V TO*xzhECAc RיI2-z,g}_{! @cKWBH#J\ݤ9Z(=TdZRPdՓX ' rkT~[>b~9j$ ]r8j}}fBܕ MBOkhOk-8D? ^u%FQV&9߇Zd6o $Ϙig=Qb.K" YOcm\'%1+:zK==׀yٌVE.Wupbܾg@의Nȷ(Ӓ_9t1I$&JSqvE<U?]J&T6oBDd hb  c $A? ?3"`?2/!a"o `!/!a"o`#@6|xڥOP=*BB VHT8 JQ#?ى 8kL\˻O{w]0@ 8ĒY1bu]Y!?LLX(` n F,+,E|wj`?'u) skT<65kR)u(e:݆&wHY24#OKa v)'d?@ABCDEFGHIJKLMNOPQRSTUVWXYZ\]^_`abehijlkmnoqpsrtuvwxy{z|}~Root Entry6 FXEgqData [FWordDocument5;ObjectPool8'xEXE_1234253003|FxExEOle CompObjfObjInfo  !"#$%&'(),/014789:;<?BEHKNORUX[^adgjmpsvy| FMicrosoft Equation 3.0 DS Equation Equation.39qigdy X=[X 1 ,X 2 ,& ,X n ]Equation Native _1234253195w FxExEOle CompObj f FMicrosoft Equation 3.0 DS Equation Equation.39qigqy x=[x 1 ,x 2 ,& ,x n ] FMicrosoft Equation 3.0 DS EqObjInfo Equation Native  _1234255366 FxExEOle CompObjfObjInfoEquation Native ]_1234255234FxExEuation Equation.39qiAdy X=(X 1 ,X 2 ) FMicrosoft Equation 3.0 DS Equation Equation.39qOle CompObjfObjInfoEquation Native iudy {p 00 ,p 10 ,p 01 ,p 11 } FMicrosoft Equation 3.0 DS Equation Equation.39qi†dy [X=g(U)_1234255563FxExEOle CompObjfObjInfoEquation Native _1234256553FxExEOle *CompObj +f]=(0,0)for0d"u<p 00 (1,0)forp 00 d"u<p 00 +p 10 (0,1)forp 00 +p 10 d"u<p 00 +p 10 +p 01 (1,1)forp 00 +p 10 +p 01 d"ud"1{ FMicrosoft Equation 3.0 DS Equation Equation.39qidy p 0 =0.05,p 1 =0.10,ObjInfo!-Equation Native ._1287947778$FxExEOle 2p 2 =0.35,p 3 =0.30p 4 =0.2 FMicrosoft Equation 3.0 DS Equation Equation.39qmH$I [X=0]~{Ud"0.05][X=1]~CompObj#%3fObjInfo&5Equation Native 6_1287948189m)FxExE[0.05<Ud"0.15][X=2]~[0.15<Ud"0.5][X=3]~[0.5<Ud"0.8][X=4]~[0.8<Ud"1] FMicrosoft Equation 3.0 DS Equation Equation.39qOle =CompObj(*>fObjInfo+@Equation Native AkOH$I F U (u)=Pr[Ud"u]=u FMicrosoft Equation 3.0 DS Equation Equation.39q'&' F X (x)_1201429240.FxExEOle CCompObj-/DfObjInfo0FEquation Native GC_12879482373FxExEOle ICompObj24Jf FMicrosoft Equation 3.0 DS Equation Equation.39q›p} F X (x)=Pr[Xd"x]=Pr[F X (X)d"F X (x)]ObjInfo5LEquation Native M_1287950620c8FxExEOle P FMicrosoft Equation 3.0 DS Equation Equation.39q'H$I F X (X) FMicrosoft Equation 3.0 DS EqCompObj79QfObjInfo:SEquation Native TC_12879485091E=FxExEOle VCompObj<>WfObjInfo?YEquation Native ZL'uation Equation.39q0H$I F X"1 (x) FMicrosoft Equation 3.0 DS Equation Equation.39q_1287948734;OBFxExEOle \CompObjAC]fObjInfoD_Equation Native `Q_1287948702GFxExEOle bCompObjFHcf58$I F X (x 2 ) FMicrosoft Equation 3.0 DS Equation Equation.39q5} F X (x 1 )ObjInfoIeEquation Native fQ_1287949335LFxExEOle h FMicrosoft Equation 3.0 DS Equation Equation.39q58$I F X (x 2 ) FMicrosoft Equation 3.0 DS EqCompObjKMifObjInfoNkEquation Native lQ_1287949372JTQFxExEOle nCompObjPRofObjInfoSqEquation Native rQuation Equation.39q5} F X (x 1 ) FMicrosoft Equation 3.0 DS Equation Equation.39q_1287949628VFxExEOle tCompObjUWufObjInfoXwEquation Native xC_1287949689@6[FxExEOle zCompObjZ\{f'H$I F X (" ) FMicrosoft Equation 3.0 DS Equation Equation.39qs+ Xd"xObjInfo]}Equation Native ~1_1287949759`FxExEOle CompObj_afObjInfobEquation Native Q_1287949845^heFxExE FMicrosoft Equation 3.0 DS Equation Equation.39q5$I {u|ud"x} FMicrosoft Equation 3.0 DS Equation Equation.39qOle CompObjdffObjInfogEquation Native yĊ {u|F X (u)d"F X (x)}'=  [F X (X)d"F X (x)] FMicrosoft Equation 3.0 DS Equation Equation.39q_1287950065jFxExEOle CompObjikfObjInfol'8$I F X (X) FMicrosoft Equation 3.0 DS Equation Equation.39q?ԑ U'=  FEquation Native C_1201429890,oFxExEOle CompObjnpfObjInfoqEquation Native [_1201429971tFxExEOle  X (X) FMicrosoft Equation 3.0 DS Equation Equation.39q?0d u'=  F X (u)CompObjsufObjInfovEquation Native [_1201430192ryFxExEOle CompObjxzfObjInfo{Equation Native C FMicrosoft Equation 3.0 DS Equation Equation.39q'HZt F X (X) FMicrosoft Equation 3.0 DS Equation Equation.39q_1201430496~FxExEOle CompObj}fObjInfo X$ P[Ud"u]=P[F X (X)d"u]=P[Xd"F X"1 (u)]=P[Xd"x]wherex'=  F X"1 (u) FMicrosoft Equation 3.0 DS EqEquation Native )_1234349057FxExEOle CompObjfuation Equation.39qiYey X~Exponential(=0.2) FMicrosoft Equation 3.0 DS Equation Equation.39qObjInfoEquation Native u_1234349122FxExEOle CompObjfObjInfoEquation Native _1234349386FxExEi©DK f X (x)=(1/)exp("x/)whereS X =[0,") FMicrosoft Equation 3.0 DS Equation Equation.39qOle CompObjfObjInfoEquation Native iT F X (x)=f X (u)du 0x +" =1"e "x/ '=  y FMicrosoft Equation 3.0 DS Equation Equation.39q_1234349470"FxExEOle CompObjfObjInfoEquation Native L_1234349592FxExEOle CompObjfi0q F X"1 (" ) FMicrosoft Equation 3.0 DS Equation Equation.39qiey e "x/ =1"y!"x/=lnObjInfoEquation Native _1287990370FxExEOle (1"y)!x="ln(1"y)!x=F X"1 (y) FMicrosoft Equation 3.0 DS Equation Equation.39q+H$I S X =CompObjfObjInfoEquation Native G_1287990756FxExE! FMicrosoft Equation 3.0 DS Equation Equation.39quH$I [Yd"y]=[" y  d"Xd" y  ]Ole CompObjfObjInfoEquation Native _1287990915FxExEOle CompObjfObjInfo FMicrosoft Equation 3.0 DS Equation Equation.39q8+ {v"[0,")|vd"y}={u"!|" y  d"ud" y  }Equation Native _1287992227FxExEOle CompObjf FMicrosoft Equation 3.0 DS Equation Equation.39q%0-$I [Yd"y] FMicrosoft Equation 3.0 DS EqObjInfoEquation Native A_1287992256FxExEOle CompObjfObjInfoEquation Native i_1287991398YFizEizEuation Equation.39qM:L [" y  d"Xd" y  ] FMicrosoft Equation 3.0 DS Equation Equation.39qOle CompObjfObjInfoEquation Native O3 9$I F Y (y)'=  Pr[Yd"y]=Pr[" y  d"Xd" y  ]=F X ( y  )"F X (" y  ) FMicrosoft Equation 3.0 DS Eq_1287991722FizEizEOle CompObjfObjInfo "%&'(),/01258;>?@CFINQRSTWZ[\]`cdehknqrstwz{|}~uation Equation.39q=H$I [" y  , y  ] FMicrosoft Equation 3.0 DS Equation Equation.39qEquation Native Y_1287991793FizEizEOle CompObjfObjInfoEquation Native 2_1287991828FizEizEOle  +  y  FMicrosoft Equation 3.0 DS Equation Equation.39q1 " y  FMicrosoft Equation 3.0 DS EqCompObj fObjInfo Equation Native  6_1287991998FizEizEOle CompObjfObjInfoEquation Native iuation Equation.39qMH$I Pr[Xd"x]a"Pr[X<x] FMicrosoft Equation 3.0 DS Equation Equation.39q5ܽߴdEVRAo*/q/_ U**wU>uxp1${8T;lݰ{4NCtVyՅ|y s?Ձ7yA@^_oK}1~~M~'\F_0o4"޿F/1?3udjeV.ɔ˟gv3Gi><ֳO}gp&0ggJeS?6DDx`N89|r.|\Zntg >SFF~|Dxtt,']D-\ <1\]ա7@ 15NakľPDd phvb  c $A? ?3"`?26E⼌$2j 3x`!6E⼌$2j 3 @|vxR=KA}ILhH%bqZ b>pb3Dvjem%"vV6))V*sfC"켙fGmCF>!}ւ}٘%Od.J 997UubmOu_MJMXvj9czŏݯ'=OY3t.hSGcҌkq\xˊ{I4o=\ xVT ~$d쵎WӼz/3+bkYw9sa|c+=Cg&Q4=+Ym;js{m<# T )+ӵA7p\ ^P]ߏ5_Њ$Dd Tb   c $A ? ?3"`?2nq|g EJ`!Bq|g E   dXJxcdd``fed``baV d,FYzP1n:&&! KA?H1Z ㆪaM,,He` @201d++&1X +|-_T TAZ5+a|i&_|Fgd6nG?y0@alBD%4Ը! h 0y{qĤ\Y\2C D,Ā'f~cX$Dd Tb   c $A ? ?3"`? 2nq|g EJ&`!Bq|g E   dXJxcdd``fed``baV d,FYzP1n:&&! KA?H1Z ㆪaM,,He` @201d++&1X +|-_T TAZ5+a|i&_|Fgd6nG?y0@alBD%4Ը! h 0y{qĤ\Y\2C D,Ā'f~cXDd xT# b   c $A ? ?3"`? 27zcg_]J`!7zcg_] ^&XJxSKQ]Q~Ht"Iz+$=Hb~ K:y.1 n=)hyn|,/p:.^nZaHu]XQd oEB_ W"nuL&8>̖rʙ yxjTüZTԲ]gO[~ī~D =\b^i9c)Jd~2hq BsvU3ûDf2]S3a.rj4;zO ̼osNn2-'_ vx~<CI&6b>gTuPefmv5Q!=4?Lén+teg' Skk> p{<4h%޾-PSdXRPT!@^}5gB ~#:Dd Tb   c $A ? ?3"`? 21.<| ,``!X1.<| ,  (+XJ&xQ1K`}%6-X(D覈v(B ![g'7NgBqΤt%w{*Wr`j%zEH$2Z^jʃ`}{i`oԙ3z_ɪyO GU;/i#/GIxzn+Pצ"Ǐ=UFx='Rd Ɖ/Z~ w| Y޷yH+x=͚d7ׄgr[5EAk* \&#_)T[{ uC%XD-1Q_ߎ!Z"wZ] _Oɉs$Dd Tb  c $A ? ?3"`?2nq|g EJ$`!Bq|g E   dXJxcdd``fed``baV d,FYzP1n:&&! KA?H1Z ㆪaM,,He` @201d++&1X +|-_T TAZ5+a|i&_|Fgd6nG?y0@alBD%4Ը! h 0y{qĤ\Y\2C D,Ā'f~cXCDd Tb  c $A ? ?3"`?2;7wƥUi&`!a;7wƥUb XJ/xcdd``6dd``baV dR`1FYzP1n:&,B@q56~) @ k/`a`847$# !lo@I A $37X/\!(?71aŪXkyB* @@ڈUg5*A~ͤtj>#Ȟp{0f15p7HwCe3ؽp1Ltp?i`\@A11 и``㘑I)$5a(: b> 0?1pxBDd Tb  c $A? ?3"`?2`eV FlGh@)`!``eV FlGb (+XJ.xcdd``6dd``baV dR`1FYzP1n:&,B@q56~) @ k'00 UXRY7S?$LY ZZǰb@`<+h(.P56j`| ҪY 3 fR H5d=BEd)wf{+wc~t18^#*b c Sq!1#RpeqIj.(: b> 0?1Xu$Dd Tb  c $A ? ?3"`?2nq|g EJ+`!Bq|g E   dXJxcdd``fed``baV d,FYzP1n:&&! KA?H1Z ㆪaM,,He` @201d++&1X +|-_T TAZ5+a|i&_|Fgd6nG?y0@alBD%4Ը! h 0y{qĤ\Y\2C D,Ā'f~cXBDd Tb  c $A? ?3"`?2AdeCUS•yh-`!`AdeCUS•yb XJ.xcdd``6dd``baV dR`1FYzP1n:&,B@q56~) @ k/`a`847$# !lo@I A $37X/\!(?71aŪXkt4U @@ڈUg5*A~ͤtj>#Ȟp{0f15p7HwCe3ؽp1Ltp?i`\@A11 и``㘑I)$5A @ ] @u1a XvBDd Tb  c $A? ?3"`?26m̋h/`!`6m̋b (+XJ.xcdd``6dd``baV dR`1FYzP1n:&,B@q56~) @ k'00 UXRY7S?$LY ZZǰb@`TP]j mUg5*A~ͤtj>#Ȟp{01p76R 1`V2 ك~cpFTrM8@2@BcF&&\P3u(2tA4|(.`~bSxrNDd Tb  c $A? ?3"`?2vS 2 &t*2`!lvS 2 &R @2XJ:xQ=K`~KhZ1dQ!v;(?!Э"kňB ݺI%CR7"A.JSSa"3%*,Rdb<7֖ZB! 8>06|lݱ`dee GK[gI 4<$ ?Gp-N"qW#5]K$~7ur9k yV>Yސgm ?]p8կLw~+̅${sMQ8A${gzl,c+u+Dd gdb  c $A? ?3"`?2uT$q7v0Qx4`!IT$q7v0 dHxcdd`` $X bbd12,(ㆫa:"Wcgb f6aP5< %! `35J,L ! ~ Ay +V}. _SֲUMF\ VT ~D%oL +ss%|a>#ȞHpq%Pc.Av1@"#RpeqIj.Z: | .F`OUNDd @b  c $A? ?3"`?27Jex]zt6`!l7Jex]z*@8 :xcdd``ed``baV dR`1FYzP1n:&&6!?P u cc`847$# !lo@I A $37X/\!(?71aŪXku,@u@@ڈ+1? v />(12d p{@R8;R {@W&00\t(ӟ{mgٳ HB  1h 0y{qĤ\Y\~ @ ]` "opPkZDd b  c $A? ?3"`?24 UtFڥ寷8`! UtFڥ寷 P%xSMKBAo>-"ZPˆ@h.jUD($W YAIImƶ XEZ~D"E{sBssfpv hq+(Jb>s8Տu^h-puf+̆08rTl|xB zb~ .78H`iцqav6't^nT=5.^+?XO8&5Η-勹8^};8bOܿ,/߯Esʰ%7X}OU;[+e ⷧc=@+ _q\yxi]Ϩ/tLba}4=kܧ)!<$luڴu e g5LD/KE|3>p0EETf3'`X@?w?8=.!?9Dd Tb  c $A? ?3"`?2шMI[yIU_;`!WшMI[yIU  (+XJ%xQ1KPKmZ0dQ :TDC7ECG;Dt -l$pv*'ݙ=}Aj؁XްUd flQ\~OЯ}dͲ=j/icovx]>${Ө~mv? FaVVLf[3[yWBa:u{ND͈9U^. trg!5[k0N~@. `!~s ,@\X Lxcdd``Ned``baV d,FYzP1n:&fv! KA?H1Z VsC0&dT20 KXB2sSRsv,\~ oLT T|@Z5+aJ_|FgdQdg 1P{g/^F&3CE7U@(\G!"ĸ 2J,b;EwoǶ\X pB\NKLLJ% @2u(2t5B ~` ^Dd @!b  c $A? ?3"`?2~za.t@`!|~za. ~ Jxcdd``Ned``baV d,FYzP1n:&f6! KA?H1Z VsC0&dT20 KXB2sSRsv,\~ LkeF\ L 0$ Qn>#I32HL[ Dž w {g/^FA&D` * Ma`utG!" vK%x@1w''nsݛd6׶ \^h 0y{iI)$5d.P"CXY`P$Dd T[b  c $A ? ?3"`?2nq|g EJB`!Bq|g E   dXJxcdd``fed``baV d,FYzP1n:&&! KA?H1Z ㆪaM,,He` @201d++&1X +|-_T TAZ5+a|i&_|Fgd6nG?y0@alBD%4Ը! h 0y{qĤ\Y\2C D,Ā'f~cX#Dd TO@b  c $A? ?3"`?2mF"fJDjpE19aWl?5CTh]Ï}i0$|Gs{.aa׳QDg~y9c~^V Cá=Z$+ssRD/(} ;x@Fb $sqmm.\F/f. ^9 1,Dd T# b ! c $A? ?3"`? 2% ;sӍo^Sz1N`! ;sӍo^Sz1 )XJxڕK@][MjkQ0vAOPvR*VHQjg]Iԩ %9l˽_}ݻ{ROmsi!*(1& 6E@!Ie6+,avy;4\X7?G`>a| 5lfTyi&<4sK n``F&&\^: @> 1 3X^Dd h>_b # c $A? ?3"`?"2~} W?NT`!|~} W?N@|Jxcdd``ndd``baV d,FYzP1n:&! KA?H1, ߁qC0&dT20 KXB2sSRsv,\~ 눦ʠ T0ndiլ[@|J 3D=8\ + +ssE:|np3ݿ{Q,u-W\Fս L`"0]2 Z) mд"AF&&\ s: @> 1,!!?,Dd hob $ c $A? ?3"`?#2vѰoxSYeRW`!JѰoxSYe/@ȢI|xcdd`` @c112BYL%bput~ ٟ6U& ml'bS=UD_AAhXчuP8))|kEh ]nmJ=EB5-җ$@YV=Psl6t-݁R/1 #c}b:_IL[PmP3u|W,5b~Gm^$-:LmjUHHk ̏ZWz. XF +~j_f;lI| [8H?L՘ ?o%X5 WIPgVIDؖɲ;db< &7>Rޗ=(̛`a<0' s9`NY]Q'M s֤͗֝[+z@[0q(nP>A^;Yw̙'\|˳*G,s;pܧ~šHr6m kEeK)"bW̾OS/dӗ-N>cd/dgi f'L2} ?4["űzq3smB[2ǿ-Bv(ۥ[%i)|c41|7ķI% ku-RX'4Y'   >.y=NiN)SX4Y)G <'\Xz`#&:cuFuFa1z I I7zdTZ_;Y:q]z6Z%xYuGMu&:buAuc.<늍S~9l/We|;c`|agXVvqQAo·M _]XnCmw$= _'h)C&ۚC 됁\^n0J\]U{fJ|u|Le\"s,\⹼ndRX NiN)u}9v\z|4)| ;2˿3oHo%9r&u={=rsMsry|kWG Vɷ%z:`dRX rWr\yuf6>G |lnmq Y+'EPT^hўU|.EJq1{1c.AKſxcc%1D阥H_1elBW1Jc.+*#?2hJpY)MlT|.M9F+S| ❖*#?+1jD-PF-T\k,sZ#F賟[9a1R_ s#Fs;8e1R7\M8j'9XlT|.MP53Ʒ;[v3]Q7H S?^IzNk4B'h׋pxGiOm\#iz#?R BNb\ >O򇴒/} K'9C0gi W<\;8 #Ź~!ǰ珧aфa6c, -X0m!UՒD#vq+e~YdWֵ;f˷s?Q(~c~b/@tDd  @b & c $A? ?3"`?%2b,#Rq͎[d`!b,#Rq͎[ p `xcdd``ed``baV d,FYzP1n:&&! KA?H1: ㆪaM,,He` @201d++&1X +|-WMT T#AZ5+a| o kO>+ssRD/(} ;x@Fb $sqmm.\F/f. ^9 1,CDd 4TbCb ' c $A ? ?3"`?&2wc oSA*irg`!awc oSA*  pXJ/xPJA};wQs $\%bEB HD3Bvil,B>JDlSߑOPUsf;+ݙ}T a=PaR"c,wǬZ%[ 9M_u67&:z󎻎c^?LD|B3rihA#<}fVl̤h1z`d<)}uE|C8mA{3ѩsێS{OzW{33y2VyM!!r{d5Two6#7 Ʋpn@tF-f9aDd &b ( c $A!? ?3"`?'2"1}]; :30 43i`!1}]; :30 43@hxSM(Da=}g`a!Y &V#Qv6,IFA鱙Ke#eAK{ kK,H,q}?L)_}u={{Q1@S7JAD)*J6? u TD=q%a\,Cm,NkY`dV*Dyism4&9#_ O"Ǯ쓢62ԨX0#u:A |h2H{oQGc_Xeדc|2[[o=ԴP3ߙ}c(N 3QCV\YlLhn66z%xHMXod }OIltX&n(2d&ˏL# +N-ӚP{_N;\(\Dkrʈ 8]$WbEZ}j+;kP#;߶qZR)5}R^*\, <|W{-0񽶚kIӼ#nI('us~T3QH <'2r[9BT?x4~ ?l-"FSA-Dt[5#3y82qWO[^h|v ۉD9uYzog0 5d ӚMDwe/Y'fADd @'b * c $A#? ?3"`?)2W6B6go`!_W6B6 (+ -xcdd``fed``baV dR`1FYzP1n:&,B@q56~) @ '00 UXRY7S?$LY ZZǰb@`@O @@ڈ+3Hkl%/GWl>Hx@|@ lO%ܞ^=1 Ma`8ϥN@a)8Wr ,8dAg;LLJ% ?Aԡ"|Eh.O`RhKDd &b + c $A$? ?3"`?*2'=a1ףxoFq`!'=a1ףxoF" FhxRJQ=w6(.H(N k@$ V a Y"b'XZH1:s! ^83s`ƹRAM1")SFj,թd8>=8q07\=-:gK8Y!/6åQP %(K'EkV3gju_v!~wVuږ(IW|]&-]`uW?HkuRo3>8G_pK.L*tUє}&q?ku2@ف+Dd 'b , c $A%? ?3"`?+2ҭʼ Jϙt`!ҭʼ Jϙ/ IxTMhA~6&1IcMlVjRIPs-bҴbXM axDث?=7E".I/ śMBy3fK{v4h:@zV (-4q뺜i,EԗFb؁NHu טW>(Lo[EhcیZ!,5&gA8CB섖 |t]W[ VYIo-%5RA<R[YGdb&(8N_ù~OKZ?: lP?8'ԷWܢ{ьo<,p@!:>[*j^+M߂o6O^;6=GC6}KwN_SO='m6XH^QWk y9D8*0U~#nW]*-aU>nx7_7ukdt=kW1abWYR143w`X_ڌ1Sg)[8NyQg}/|T_TU`]u-L^Y ۽Qhhbߘ]J0E ^x_XيC: Dd CSb - c $A&? ?3"`?,2)E|.3! 8Zǰ^x`!)E|.3! 8Zǰ@0= hxRKQ ݵZh)xJq"th` -C7'MD'Ap(.]?*ݏ`K #TF/BA)躮DKbڏ%C^ X_adzb/JIZWu,meMW$iMť`c*I6+|1k u:DB֩8N->?B 绕/::tyQN"{7(+Ni0}U,P90;qnz=Dd }}b . c $A'? ?3"`?-2M۸a[eƌc{`![M۸a[eƌhh)x]PK`}w?V CAPE? B %ܜqsqM(.w%ݽr5c]DqYݡUWp Ӕz{~kl r/${8Da8/:M bQ?U_b[~Ӭ#Z-3l,==Nʊ?^6U,[Lxm3et'Y /D=\:f`m/(̘y:#% XHy$ P=cFP/Gų1<,F~aDd 0$-b / c $A(? ?3"`?.2H۪w;-WmEB}`!H۪w;-WmEkhMxuQ=KA}3w~͑X8%EB,HRN<#$]ZU Kl ٻSe͛7; `-`cj3N#&I8"mrey9q*1 G=W-`Gل~s/}~JV+l,FE\OK F1/[̯I=Ө5T-{S#[3pgWk <L7 F X (x)=.005x+0.5ObjInfoEquation Native _1287993821FizEizEOle   FMicrosoft Equation 3.0 DS Equation Equation.39q\)$I F Y (y)=F X ( y  )"F X (" y  )=(.005 y  +0.5)CompObj!fObjInfo#Equation Native $x_1287994081FizEizE"(".005 y  +0.5)=0.01 y  FMicrosoft Equation 3.0 DS Equation Equation.39qH$I f Y (y)=ddyF Y (yOle *CompObj+fObjInfo-Equation Native .)=ddy0.01 y  =0.005 y  FMicrosoft Equation 3.0 DS Equation Equation.39q_H$I 'lim y!_1287994965FizEizEOle 3CompObj4fObjInfo6      !"#$%s&()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrtwxyz{|}~Equation Native 7{_1288013523&FizEizEOle 9CompObj:f0 f Y (y)=" FMicrosoft Equation 3.0 DS Equation Equation.39q®2$I F Y (y)=u=0.01 y  !y=(100u) ObjInfo<Equation Native =_12880140490FizEizEOle A2 FMicrosoft Equation 3.0 DS Equation Equation.39qEH$I X=200U"100CompObjBfObjInfoDEquation Native Ea_1287999364FizEizEOle GCompObjHfObjInfoJEquation Native K; FMicrosoft Equation 3.0 DS Equation Equation.39qH$I  X2 FMicrosoft Equation 3.0 DS Equation Equation.39q_1288000949FizEizEOle LCompObjMfObjInfoOEquation Native P7_1288000306FizEizEOle UCompObjVf6{ F X (x)=f X (u)du= ""x +" 1 X  2  e "u 2 /2 X2 du ""x +" FMicrosoft Equation 3.0 DS Equation Equation.39qObjInfoXEquation Native Y_1288000473FizEizEOle ^ 9$I f Y (y)=ddyF Y (y)=ddyF X ( y  )"ddyF X (" y  ) FMicrosoft Equation 3.0 DS EqCompObj_fObjInfoaEquation Native b_1288000680FizEizEuation Equation.39q²P/$I ddyF X ( y  )=dF X (g)dg" dg(y)dy FMicrosoft Equation 3.0 DS EqOle fCompObjgfObjInfoiEquation Native j^uation Equation.39qBqy g(y)'=   y  FMicrosoft Equation 3.0 DS Equation Equation.39q_1288000850 FizEizEOle lCompObj  mfObjInfo oEquation Native p_1288002783FizEizEOle uCompObjvf 2$I dF X (g)dg=f X (g)anddg(y)dy=ddyy 1/2 =12y "1/2 FMicrosoft Equation 3.0 DS Equation Equation.39qObjInfoxEquation Native y_1288011603 FizEizEOle Ž 2$I ddyF X ( y  )=dF X (g)dg" dg(y)dy=f X ( y  )" 12 y  =12 2 X2 y  e "y/2 X2 FMicrosoft Equation 3.0 DS Equation Equation.39q2] ddyF X (" y  )="f X ( y  )" 12 y  =12 2 XCompObjfObjInfoEquation Native N_1288011761FizEizE2 y  e "y/2 X2 FMicrosoft Equation 3.0 DS Equation Equation.39q2$I f Y (y)=1 2 X2 yOle CompObjfObjInfoEquation Native   e "y/2 X2 forye"0 FMicrosoft Equation 3.0 DS Equation Equation.39qG8$I W'=  Y/ X2_1288012251FizEizEOle CompObjfObjInfo Equation Native c_1288012449#FizEizEOle CompObj"$f FMicrosoft Equation 3.0 DS Equation Equation.39qH$I F W (w)=Pr[Wd"w]=Pr[Yd" X2 w]=F Y ( X2 w)ObjInfo%Equation Native _1288012685!+(FizEizEOle  FMicrosoft Equation 3.0 DS Equation Equation.39qH$I f W (w)= X2 " f Y ( X2 w)=1 2w  e "w/CompObj')fObjInfo*Equation Native _1288012931-FizEizE2 FMicrosoft Equation 3.0 DS Equation Equation.39q±*Ԇ W'=  Y/ X2 a"(X/ X ) 2 '=  ZOle CompObj,.fObjInfo/Equation Native  2 FMicrosoft Equation 3.0 DS Equation Equation.39q;H$I 3 X =100Oh+'0_12880692842FizEizEOle CompObj13fObjInfo4Equation Native W1Tablev5SummaryInformation(7DocumentSummaryInformation8H/2f DeW4T`!f DeW4Tn p \xRJ@}II 9xUD b{DBDXI17Ϟыzӛڃ{'qf*.f̛30")6djIA!2ce͉WxE gG,L1`~$KtY!hԣ_OAZ\]jov?+{DQ.iqFDRȪ^ʸ)F22 uH8p 2+r$EVsx7^ t'c|RV<>wg\ly )gjQ( @8Q |tzE'c;spDd Tb 1 c $A*? ?3"`?02袚v9聯2 `!袚v9聯2  (XJ\xR=KP=%&-H ~Dg΢"K) ZHGW77'@8`5P{n;w"`ږ$Umʬ>IjELb>WMf@ `~vΔSԢsSڟ2, ]®~Cwps;h{85l4LDu+ƕig`U/k9nis\<~GU^lHU3^C^+긔:Œb8,b-1k!MƘBRUN@'fq9<@sGG݉=yCDd Tb 2 c $A+? ?3"`?12Yק|1ОF`!Yק|1ОF xXJuxRJ@}mMƞDP㛑yKXNϟi҅iQHBȅ 6!E6I.mQ6imjSq7G-p|;;{{k@6(!rZ9*ג/E[tzׁ ̷fxY/8''nEKס1wHBc6YLl2yh+_D.2-7Ʌ0FR*sY?)9G%[neP]iSmqyytMruuɅoZB_ߐG|>~^ .0aRSN\G%WBr! tXϻ/A!=6U.ͺJ7)OLʓs6We=99PG}*cF 7yMbiCzb&^eo~c&\.]nFK2/yԱ+6SG6t=SzU=_GAu~/5?qQ~㠗}}]b$bIE;w ??5ҁ"\@Ek33Ƒ3jg&L+xWQ-:O{d|(`s:nħ4y%{4@LGL=+c_z*瀴=4 0$/|҉aS6EBBk.(Gv OvEx>9V[4u}j'#+$9φ(%!ϻ6+{![b^Y' tsJ}=>co O:44.enYs_H in[l`b@8M(MT E3] k(Dd |b 5 c $A.? ?3"`?42ک c)?ZYjc`!ک c)?ZYjD  xtxR=KA}2T"DD b¤Fۥ [-DE,D; /OPUsfswAc1#ݨ")IׄtGG+bʭ|\\v$i/'1 :@6Et1c.&b ,Qkd霱 J$s_-)c#n㉮l\J  MX5<_oÜ+֓&g?2F λS`GvŪ[VܮV# .9(EebÎvwM;P7Ɍ%Wg0t~fuu rGOHUy' mIDt[,˩Zݪ 1µimx7tۦ>jDd 0 6 # A/5"𓴃E@=𓴃Ei1cQ =x UUr&:DbJ\HS %R.MPn.S!Esٟw2kTv]:zڹAP؂"v !2!B.LXPkal"̷``"gby.}1< c*sa0w;sa0w;(Gā<lK V"@iVdp# 3ѼbNYr7=FSe|cy{ YiX2{U]ák\7WMqCk-瀑 [t|_ёCήbcf}mc>5tbwڷ_HwNSFg:so p}LK{@Im@P}wM&iM$6rҤc: ɟ ɼ$|pLy84#-8UԥRp(!=MU"u7NUŷtKbJk&cNڗŪ:իWW ڲ:cl{k _-t|ΫK[*~__㷜qv8p qٯ4656XJf':l'IT&%=jds>$kqdgdb1Y,1žٸĜ>cx_sP-9cbR9ȵqh{'>h[u~ENPρUmvTps"ƥd(+V:鱔o,>{Z's1.HSYVxO"-V:i!X=e>1Ƹ AL[R(Lok+MfI2{ʇdVaLuzˇ޲meV`\Qz˗^z[[izا޲me[>|}Xo+Ĵ [weNX4'w飷z[glGoهOLYQz˗>>2I;3e@|-V5>+Jo޽ HQ4[$Xv棷z[HZo[a`|V%z[[izދٙ޲me iGoهOLsYQz˗^N^HNFmm2Iײ3e@zˇ޲me YQz˗^O^HNFmm}tz=d>9e>1}ϊ{AȗBr֖BM-h-lض.JkWKs3}M VfkĴA deA/F[u zOU0瀵.Ȧ- (=U_x/L'&>10HV߾~ }s!iC[;Yq:-DrPc0{۶Sn eZe>1IpLVIXc1ڏTOlښ\Kl5}LI뛅k㍥ǛsAT&%}kb Ϲ8i1Y98u12 HvQN:h@mJk[>6"yIk=.PCx/`Sp5|-|?V1h~m~_WNw{]%زC ۹4I0_'JmM2N].V&U ͢85Uuq[k]ǯZ_C\O`Q Ԧ:0w11SC̑HnǛs>(kO9I^z3}>1c[Y|hμYMKP\6 ٿ%ζS pLjIe+Gm[ca1Fskm~xˎrJWd1ߒ{{oM Vf䜏|&:i'ITOLq9cdeqv1ٸ$3No6RBH1|j12UI/ xj4ڋ?cWy5_y!E?M";_r06.YE9GdT'=᣷@M瀕L:W%mmٗا޲me[[aӓW%!Jgݬۈ9NZo[a&+Joޕ!Vok+Md>}}Xo+|-Vio|V%X4}5á&Y3᣷z['(KzX4}=z[glGoهOLYQz˗ӀRJgݬދ 6}V>|}Xo+4|IϺ׾֤ʇ޲me YQz˗ֺX4K[7[{/#^kd_6Xͤu'm}bjqs$@Jxk1%?{IN/v%&}Ӯsl;%tMN4'&i>1iY ƩGK%TH ˪ 8Wa}U&y de4'' _7N}զD63}eWe>1IpLVI`ʫo|IߦVڏTH洗U_}gWa}U&y de4i Ʃ//9?MumA{YA_}eWe>1IpLVIjSy݊;[X˰j+Mߖ [C|m/@1X[jƺ8W8T/1h.m7Euԕx>򿫿)8b"Uߓ9B Z)߯6;>l4~oZ][L6߶kSTaMI,eѤ}}ƚ&zS^1_jҡVWEhևYgZ]}ٰyiTR BJ;dzikG֋cILC6W> .{\2UI.)}'1786NVN4Nī <Z C㌁r:}t^h>ʕ,׶|FSmDrs'<ƺ$\!hB2xb@Ot <Ѕ4@Woֺ@7a)ɪ7+Ogp&blɞaO D@|V0.\#`$\G0X0.KF:`"L`2Lቂp9L+`:̀+*6f, 00TX p#fn ]p7K^a,xQx xVSPOC!ܰ9ɼн ?Jx`/˰'/^5^uxބmx6G1|g>/`|ɓ_foa |~`lvooA2T B5{@u{P+{Cm;ԁ@B]?_Pp4 hCPh4fp4#-hhp!ln؜(`^(jo'Nvp2{:B' ]+dC7пziN3L8 @_g9p A0·!0 aQp!10Ep1KRy0&e0T0fp\ 3ăkZ\a, p b;a w= r!xG1x'a<\9͍WPX/K2W` W5xހ7-xށu.z> | l _Wo[l'g~`Ucw^ל]#%* B%Qi{YwO{F'G=s:4BҙR~Tn'vn3:'#tMF jSخm*G,ڒu y*,嵓l7~I]^&l~ $$b f;-?Ꞃe7־׉X8/Xm ][oi# {czMSFTzK,~/q|t[mզ9J0[mզ9pNTzK,{ [ڗNn慄EoVk+1ToV9Xlzk_z2o6́Vb!\mM=}Ius-j o뭶KNl*%/־z;#颷zEoV%CTzK,{ L~(ڗNL梷z#Lz0b6K~/EoVem%o뭶KNӉTzK,.}>89][o]Vb-j䴜M_ڗ^`w-s-jp[mvi1B1$!Do ;AuI߹BFYI>c0fuQU_O 4L<6%5']r:q))@IH:m[~(p'mɎ|r: Wu-E_)N$z9풓1)֩$ϖ}ORۅۉ`.'}]r=vI?:5V߽%凲E_银T@C"o+6I_DO;']r}[X+C٢4H[*}~X}Wl.9vNj$~ )@iWە1 շTWl.9vNj$> mT4Gs}W1^җN߃P}2Vs[_)N$z9풓t=C_cxѷJ_:}1Ncl=o+6I_DO;']r}]W}}g__Kow4TߑMWm}Ŧ8뒓i礶KNXpշXYWۃ1 (&[S+bSuIsR%'Wzd:^#{ tr''D}quR+bSuIsR%8zW,w>>Ly$aVK/޻[0bD%B;C2h.:N4籲/2V?"Jۏ8/en~緤t 4y<Ӷҧz7g6UC]< ǦgF!Jwל:?O'&o'MA4igɆ7&>h19>K֑+Og9ڗ3,׏%ql?" [ɮMDڤ6MMS qu 9.9~&;')mT{IT93r!aAcDx=uJѭ%fi%Lێ%¶--ͥh}{X7?'@ 2a0'+)p* Bs%48pgHL 5BQ\k.K.f{3=Eֳt6&d1g3fC9D\FΥg.ؿuf\7p-7C9̇[na;a!w"*`)~xAX<ʌ+`%< T2zւzLZ86b< cEx ^WW5xހ7mxxއC>f>O3/Y_?V[6n/7Ա[.2͠9d@vLȂWK 5z6kh ͫ0~ =!-5{A{=tw/a>'ȅ.p] ~ C8?__P8 '#eHzl1c*qlHQp44}B>xo(Ap" 2'p Cita#a1g ,(a9\( 0&y s>L BQ.K.pL {\Wte0kaWif <n&oa>mp;,;͝a,%pTxRXrx{< VcP qXU' @A51!#*$cB5<O3P/ޓx^ xނal֚w]xއC>OS|%|7u7ˡ1vcEWbuk{5_Ɇv8*ǜD\6iM^aC)h%wΧoO}n2^3_|'z(}VJŠIB$;bw߳"s||NoJ}g%n rLK A>cďp-mcLld:'&[]bn|tg||4Ο)b wn+;{xY!ɺj=D:Q>b[B󃷰~\רBXQ{gaabq/wԲ*aDA?Dd &b 8 c $A1? ?3"`?72Ct=vYFn<<)[WĶ`!t=vYFn<<)[W)hxS+DQ?ax !&7|Dofa# V,l|DRRJv씬,fom?dcs})n:s~w9^5  YǑ!^#%@\ ND$0Jp^u$nYa;Dglx)j# ǁg<;n&Q?"N۹Rx3LpfAqO>㩈UbrrrDr*b״e@ t"w͛J ]MDd t @9'b 9 c $A2? ?3"`?82OBOs`!kOBO q 9xJAKq)dM#(b"4Šg Y[b>UjGP9]pagۙUh&dٚTzj<ޮZ/V(CwV|lss 90b"`rF8/C^E3M -ҟ /5΋9@=Sޞ֦TNtX$n=|0άUǬ̕gƺɑx_/sDO/Uq pRq]eڮh(aE-3 lbGybMDd t @9'b : c $A2? ?3"`?92OBOs `!kOBO q 9xJAKq)dM#(b"4Šg Y[b>UjGP9]pagۙUh&dٚTzj<ޮZ/V(CwV|lss 90b"`rF8/C^E3M -ҟ /5΋9@=Sޞ֦TNtX$n=|0άUǬ̕gƺɑx_/sDO/Uq pRq]eڮh(aE-3 lbGybADd 0 ; # A3:"YYҙ;j|W@=YYҙ;j|0d1cQ =_xlUS{i8 "$lVW3R@ Թ ԡY좑-%1e[DBLf₫&#GDX`3\F悮|ϽOy}s􈺾y}}㉜s?s@3ܮJtqs+_\78X;1@ss]P/nt]gᖻZ6;]_y Ӏ Z2dRe|JOElKen7+or n԰mWL{(*7r[jr=:VPn**ou߷]at_¸a ilʩwHeҮ%_tȵcǖ~օQ_h=4w_.7yV?}M:Mz4~ΐiM?,MvNo: Dө NGvLcfelǖM~:Iw'c6溹Eĥٓα6cwh.H_ʓKPXIcMޯ2>6溹E`ǎٓ1p(5~mUӻNnm,j"-_okgisNޚKz_Bޝu}+kf[­Ȣ}%1D魹] 齰Na,zYYS7/Po+0.Ke޿~[΍IԦ}E&k=YbsNK뷦Pw뫮ݹ}-ԝexLʁ%r// kg[eDbv|&eg3u;s];lG9X>ْځx*&qOSƥ1g:Ƴ$=%&~yjR撾W唾ꓭӟVF&ˢ}}զd7KLӏYbS+T1=9KtW}UwYڔLfIz1Y;KLeS\U_5}ĘIq={ 6%Yb~L}p;luõ ' I Xø^I_Mɤo$}Z]\w-UlϗҟV"c6{%}6%E&C.͞dmLu$}T/}/ҿR_5};[wn)Aջc 8 >X_#!W~hJτ_s،Ԟ3vGXsևc[?/Qyrk~8z5DtQ8 S;Ofpi|ϋ0>qƧvW‘> 9j/`=#yq'Γ# 5P;XtQ8 S;O^cpk7c[?/Qyrt5QXvk!ҭGa|(O<9jG]#yq'Γ#c=fhU4?F(O'G^ .8Z e1ҭGa|(O<9Z7|5EEc[?/Qyrnz"H~^0>h=ܬzh=kVGwH~^0>hs<6f/F(O'G[mn8 Ga|j'qd 'JonTkZ%Yom^~vg<=+-=꺣7l>w(ӣk\u;w{^Onp!x3 U9cYWGm?5՛=cCͮ4@>tr``}-Qf g,6YIp[TV)!畛OVM緌[}}[%M;տKLokүIE.4?`4ݩg>H-&^|[^v_wr֤.JtAC)M nMe3mCEl_S5q Ġfjqi=eMn_o1&if?VZ'U9}c91UʱM#ךc<{5;y5 N9fJ9և_|?3hO=Ͷ~un>{ʣ0{n q_B&6Dd hh''b < c $A4? ?3"`?;2nbQGFŽ=\`!TnbQGFŽ=@@||"xP=KA}3")N 1 V.(x&pc;kt)&?IB7Fpy!T!X]ʼn~, 7-STu >73{%ײƊ]z1`sG 52H69gbͲ͚w.%)?bwVa:{S-z(Y78^Dd B9b = c $A5? ?3"`?<2 vW͑ G!񉻳 `!vW͑ G!񉻳 `@+`\xU=LSQ>K)ã- єpp2:c"B"C,@mJ7Gcbt12,FC &ry=hMn{s;Sg4~B *em ժS=u`xIuBo!\c.4֧.NndPq٢hs#G$j&$RxOI.Y>$߷̦j{e`ыnuǏj7e+ sA 8ܵY]~$sߥlԊύ%.htӟ/؏RE&kkѿϜ%73hb| c $A6? ?3"`?=27tb(J_*O.`!{7tb(J_*O. 81dIxT=hQݻr?%w!DDTʥN Р%x!S8*DEp bg[jVL]g{{K|73;3 `>@ FWv0e`ǂ=e *_ׇV] 0  d^2*.--w_)+xY4wӏ3)Jd}M'+/}ʑ2n9INdNCQΟ5uC9p!~fySqa1tuLA9]/O)]2KjN?6?'Ts2y~^Ssq㟢>%Ac66ۯmoםRu>t$˯-^x=}Axl6X]k{8 On Έ$Z@@#Dd 71b ? c $A7? ?3"`?>2mKT>9BٽI`!AKT>9Bٽ xdxSKQ7oNkT&Eؤ Xij;*`s@C@ -Kea)Dɉ̼[7of~@kp6*e%fS}eNH"@M:I;tWx'Óƣ%BGq辤\MKk\uX w8> ls0x}):0|&i~cݲ5pJ œP-R5/Oͭ6ݠoOzP8pki-FӀ[Q ΋}kzɽ+Ig22[JKp?2'nu`{124W>OLTrݷ:B:w4nyAr:Bv\qjen][]K:Weyu%s.%G+C$懲n;3WC{Cü<"t' 3>`_⼍ӃqW8}\_v ilXgޏd=gOOӃRuQ7n5~O]{1F:HivWRA*~h`C-l`$jrB OںS2\'K ~_Z-Dd 70b B c $A:? ?3"`?A2Zzb5]`!Zzb5].8H(+vxVKTao嶻zMKf "' %D"NjŮ. >ORA= Bw B " 03Hj9k 偅o7ss2, ʀ.塱1cN%a!_+Qk0B( <%GuuL8w-YjAzQ "`zÔo `)>L~XxRLW;qkOcGq% + 9\xe=_c|~ۃwCW>f57ř}DU^;_}1gr֝J'NP5;.sʝc%tE!µ'اN>|N>rV;w68y3Ko<>krl{]B )}Q9vp: ~ƽN<>y=2ƽI!#MwNlb7ZDKυ8܍^]#nü1V[@zݚ1.?*^G_5HYac@|扴'@Z/34u]pxe͜ O$q~/d]RKW>nf6N|f>GJ ?XĞ)%*{E}kO>!$zL 5a 269@Gz RpB1Dd 70b C c $A;? ?3"`?B2{v$EXIGW`!Ov$EXIG%Ȅ:(+xUOHA[u] jBCJtXcwpua%y.E^KuN:DR/nY[]H>x{{o|à /s@%*2| ͶCW]c'*#h a#2Ĥ,&]d؀O\z~HxhLo2vڋCqpBjkU{tT?k_ĸVaBUZVO[~~Hc,)N{_`W@jiF)59? }S|H.jdsF`]Pkrr4|m`L `L%2XHj`^\{%Pؤd]Nw}k Qgdi,{,ue,)Kt8&տuzVW`q4>Lc?6#1ϵOF=p_/y?}ԷOE'_Y,'/} ⷰ^cot}1L7BBu@1׉?',eɷ!K@۸i ﴿H}R7FZ|3}e;#`@!gAB!%j3b̜e:~Y\R'O C ȷ'MgN 1>Yz) =G7 ^r#V_ʯC'0CV-'ZSim+kV{9pp>˱.WW,XWwnz g!WƇ&mqovH]6͸bp(&lƗ v,S%}yTzֈC:ɋWzb&f|S qc"wC&q~va-V_,=; 1ϧ'_= QD/EIV-x;[=IaO0Z>=yIļeU?ޤM1:%2ZT,T=w|ձu{/u޼zveU ԰n<m5Y%PSRhKMTxv1z>`!;D~ HtlZ  vxROPA cC5 I&VL !\\D6Pℱ= */w_B0@X)8=RJ55$ciίFg&{5Vq}"IbX9{pgw92pTÒ[Sfo|UoSgezҚ[tI=-P."ON³M|C|' pR9H|9=-]ߖzH0}꽔C{aC+xQR$|wC|Gqx\7f 㙜Z2 -tVD9\q\Eo%y4:i=#ø}q Dd |b F c $A>? ?3"`?E2Sۇna_8MDh^\/`!'ۇna_8MDh^\f`P.0xSK[QIK4MJtvh)PdS VJŔ>Q0*ֵES CC NN N޻E}py9!|mb)5 ;RY%h,C6>vyg1 L6fuXڗI~({f蔳N%;V,YeWEek帗9d$`wrkG'،Rq/5JA[%ܷ|AAz9R^st>9_]t^}tl4-*yǻ|RYagYj>/9{~"گ.hW*%MOY!?I1?m?uk@ܣ|CC:v~3:m5$|!kQ8KS+gFF{z[p3K฽~2'2;5fQDd Hb G c $A?? ?3"`?F2'g  gw`!o'g  g6@X-=xTkA~fkM4mRښ**PA(PQ1AsA<^D<^=׶Zh|ofvMC7{1;28^= a"fSUωaKY^1gd`S% e8i2xo6}Gx/"I4Y|"GxW^܀24Wn;r>.㞊GG~}ŀëΆ/jX Ks]@_pw@qCk#G"['+|cVZiuחʟMѐCCu*bGfx\?k4qdP}l7>_BZ5/?7--*ZݿjQ݆ 3,LVs uWXY4s&3QiY?ԜO&z}盲q\Ɵ-~ڣХ!{Z:GYj^]+Ws޶]W'gbX _vgj$ޝm߼--80Dd b H c $A@? ?3"`?G2"OGgG"`!OGgG"xAKQZjږ:ԡk)huY SPHNաcN/7t2ͼ]!䃷3f+ CGZJ`ד_WŜm+ :ٗ'B0nѪAh Nz1=bbx\3YPjR13W o/ڋF84i{hX',͹v^4 `n[q ܔkm9kJhJosus;SálYX(Y:Ǡl9hon A=CdN,tMG|Q:s|Pκ:9O!<4RuK{Y*!$e7Wof4jNi/28oTLi40)US@\JjuEhvtL]2Y*WDd Tub I c $AA? ?3"`?H2{.QT'7C}`!u{.QT'7CT %XJCxQJA};9w98 X"O0I`D3k$U>BDSXY'(JYD,00;7 6PParf8l*2Rdj,3YM-/%MVߟ9/LV1.ZfzFȧ'^< c]}ЎZxx%O:HӪŸMFv\l+B3l'&HitҾwQzcݜWrzƯi:wIx~e|eGّh!wìVj#;a8 GD/fs9Dd 0 J # ABI")̪Y4^sEo= Q@=)̪Y4^sEo= 1cQ =Wxy`'@$*j[RD@T@AY%(hPqAeSAE *(-J\J[,&s |'ɛޛ{GOe2;睙{oNBF7A{8Z5 ;2ԗeOozM3ਖAa$kLiyؖ[y5z]Aw N . ヱ}g `oȆA6>mtdлAaq~e+@XvG::԰APZ~ro;5a]#֕Y22O=e'[?>I=ef0mz,tб~cȄ@_]Cu}\cRC#qYC˔A3-ÁT?&ho1~@mߋiR$Szf\jX2 x>;OM# Mq&\'ҥ{pNGr9phJ}= q4v[] ݝ!.kNRl&[zmUl&#j?bճNeݺu_U}!j,c4v{kuͅa,츶_ԯ̓|ۖyP-ύl=46Abm6':lųOݘԦOL{sƃڔm5(Ƥ6%xLx10{'|S]lxT|Mp=ʺOn9Q[ !xq㣯]}Mws)śKw}bt;?%}BT_kX<}{3%]ln]>zt|-ڀlӥ|I^1}Nynh\m[i[mc}\z>zt|-{WokĴt-_һyƿmc)E>z>s+>z>1g/݁m>>E{pLemmemmNg/ݟmTokL)]1}}V>|}b[4Qm1OHQGoٻz[gnGoٻz['L%mc%E9޲wʇ޲wOL/3]z˗65)Ꝏj]OmJʯs#p ~L|C!_[zX&-yZ?:\-Vw|ڃth mTr>9vcOLJecWJX8Rʺ S_(d,œ]q}/]}զxtcOL1 6yJV!@ 6ւxf,7)[KRz[m>1I_7&k$rW%KNNvykz[S[?+mmG΁s<a`CB"FHa Ka 0@_0p%T k`\ pL&fLam0p .XC0,E< a <Ka, xVS4<+YxUJxK2k`a9q*e2kaC9o[6w`. 6fxއCǰ')|P ۡ vP 1;/K vn/@1)ȀfZ@&-! }`_h5dpmp|r`.h ?p?#=? X8:N.p" ~_¯$ȃp2N`a9-Cݠ; ( ΄pΆp | 00F( @c`,\ `w=P}p?< AxGa1,`),8<O x g`%< VC)/x ^W@%XN9 }_5(a=&o'x6/W`l}>-| [a|gP`|N`{ƶtr k);Zɮݣ )xOn<C8'P= ,} x | 00F( @c`,\ `/[a86m89p0|C{}ap8Ñ~GBȅp,x '@8~?_/WpAW8~ oTܰ{rC㿅|(Ӡtp:z΂>΁s<a`CB"FHa Ka 0`\dWP S*ip-\ 0nf3ă9p \p'̇nJ^`<ð#(,%,e'IXO g9xVj(EX/ (7nl؞Mv7Gҟz5>Cӳ~ut\ ݝ!lky-bȦ Vu\__Xz&uJ.DcX4&6W;׼ȿ:Yyk[WsxOmJnh75eܘԦөgqSwR7nnƻ>Tј4Yx!|ͼ4Oq㣯]}Mws) \Z{]}ӥ|IÉӷ#[~>t|+{W_ktWXlw-+{W_kӝᣯ]}SWvjET_kLMFߝ磯]}MwʇwOL:t+_+kc dM+{W_ks}ekm~t+_Q}mg<}{>>s)>>1t+_ҷmT_c(OF"+{W_k̥|+{W_k4Y`ݙb8o#kc %ƓWᣯ]}΅g9sIA>|}bWlC x2|}R>|}bZt+_-Q}mx2QW_S+>>1}|I >`kc;J!}]eؽ}|ٶkS'̧-&kĤߧdS~W5@{UԗH\]xg}m͝}o|l.MiǎΥNjIzb6])?;Scɼ~UpwܛNWJ_Z-}ݘtx ҥ} ƒ;۴1M)]]ߊAnLjSR|ڱ׷'=a4nLN4Õ\Ȅ|`>{_}C}@4׵UYWjR7}]9@y`AOOVa|V``VcY%?UE0Z~L+ƚckV1VXOc U5S1IXcr1$Xcq1XOc q514ŽۡSs#$K_QŶ.7/ȍȍ?grƻ&rc3>!!Bn|Dn|Dn|Ln|Lnl%7[ɍm6rc$x6a3lmP;a@ȍȍȍ rܨ 7*ȍ rܨ 7*ȍ rܨ$7*ɍJrܨ$7*ɍJrܨ$7*ɍJrܨ$7ɍL0|(rX+a5ɨ"7ȍ*rܨ"7ȍ*rܨ"7ȍ*r M$݃إ߯<3Y=s@&^Op 1ٟGC`O-]bBn鼶fbɦ Vu__ۘ?Y=z_ȅhLƤFj:g?9|7|>fǵ~| Vu;_b8n{ ͽ4hM_6Ք:iǎ'InLjd9fN%=m,gl3O{v1ټ2O~~z7wk(T]GGbR}_ u$yvIҳzuu76{pwp>տ3>zt<᣷]0ӥ|IF-π>|NYo[֦;`.Kz}FToKNﺿbZk{>Z|NynGoٻz['UKoF>׷r-{Wokӝ᣷]Vӥ|Iom6u6.emmemm3]z˗Qm,9o^>z>s+>z>1t-_һ?Dd8޲wʇ޲wOL3]z˗(ul,9Ohp0emmemmng/ zXrzwiGoٻz[gnGoٻz['.Kzo6'6Y]Goٻz[gnGoٻz['.KzlzX2zk5uR0|[z%~MwͤOL;-&kĤlzM`sƓl7|2u6x2 rʤ~c }+{K6K_}bҼJ0OML|I O}5Dx31}e6K_}bM0OMglGS_"};1GG_ٻM'&dmoT}×Ȗv֘ۙqWd?q<=!; k`c-w+ :!>Ud,aPv η&lˮhtٓǍX_;P9{rդAAO-3Ql}Xߡ:vtps |'#e^V }sՎۖJ}bQ냜YCA3-}{TiF]}~o,4Vߏg|;ڭ@&_4YI~냋\Mwi.M\}Ǵʤnt pְCIc?vlGOLkckגcCLG>?=vQWsI1M[ZW1ލIm̧;y2mh1 mAS+|c"DG:i7׵UYWjR5 etY1c#?Za0 x_W0f-0{a,%ga5Zx]x6a'hq\Fl6 ߃pt}<8 >`# qPS`*L`́0JXa :l!l ;t^KRޑ(Gg^u9#!9#!9#!9#!9#!9#!9#!9#!9#!9#!9#!9#!9#!9#!9#!9#!9#!9#!9#!9!9#!9#!9#!9#!9#!9#!9#!9"Cr$$GBr$$G8W=08()0 0\%,Ű VJXkUxmx6fAC#;z5vÖFe,yIݞ++нp^ov$w(e]{m?5J!]xoA'vg *Iq kJ16J`)&N,58 yP b %J6v|;0/Ǽ󥘗`>bCa^y湘<[&sΠLٖR(P 3   4@ ` l x0Simulation of Measurements of a Random VariablePeter Sherman Normal.dotm shermanp6Microsoft Office Word@Z7@ FEE@=E /<՜.+,0 hp|  $F 0Simulation of Measurements of a Random Variable Title  F'Microsoft Office Word 97-2003 Document MSWordDocWord.Document.89q^ 02 0@P`p2( 0@P`p 0@P`p 0@P`p 0@P`p 0@P`p 0@P`p8XV~_HmH nH sH tH 8`8 Normal_HmH sH tH DA`D Default Paragraph FontViV  Table Normal :V 44 la (k (No List 4@4 d0Header  !.)@. d Page Number4 4 hCFooter  H$.!. hC Footer Char.1. hC0 Header CharPK![Content_Types].xmlj0Eжr(΢Iw},-j4 wP-t#bΙ{UTU^hd}㨫)*1P' ^W0)T9<l#$yi};~@(Hu* Dנz/0ǰ $ X3aZ,D0j~3߶b~i>3\`?/[G\!-Rk.sԻ..a濭?PK!֧6 _rels/.relsj0 }Q%v/C/}(h"O = C?hv=Ʌ%[xp{۵_Pѣ<1H0ORBdJE4b$q_6LR7`0̞O,En7Lib/SeеPK!kytheme/theme/themeManager.xml M @}w7c(EbˮCAǠҟ7՛K Y, e.|,H,lxɴIsQ}#Ր ֵ+!,^$j=GW)E+& 8PK!Ptheme/theme/theme1.xmlYOo6w toc'vuر-MniP@I}úama[إ4:lЯGRX^6؊>$ !)O^rC$y@/yH*񄴽)޵߻UDb`}"qۋJחX^)I`nEp)liV[]1M<OP6r=zgbIguSebORD۫qu gZo~ٺlAplxpT0+[}`jzAV2Fi@qv֬5\|ʜ̭NleXdsjcs7f W+Ն7`g ȘJj|h(KD- dXiJ؇(x$( :;˹! I_TS 1?E??ZBΪmU/?~xY'y5g&΋/ɋ>GMGeD3Vq%'#q$8K)fw9:ĵ x}rxwr:\TZaG*y8IjbRc|XŻǿI u3KGnD1NIBs RuK>V.EL+M2#'fi ~V vl{u8zH *:(W☕ ~JTe\O*tHGHY}KNP*ݾ˦TѼ9/#A7qZ$*c?qUnwN%Oi4 =3ڗP 1Pm \\9Mؓ2aD];Yt\[x]}Wr|]g- eW )6-rCSj id DЇAΜIqbJ#x꺃 6k#ASh&ʌt(Q%p%m&]caSl=X\P1Mh9MVdDAaVB[݈fJíP|8 քAV^f Hn- "d>znNJ ة>b&2vKyϼD:,AGm\nziÙ.uχYC6OMf3or$5NHT[XF64T,ќM0E)`#5XY`פ;%1U٥m;R>QD DcpU'&LE/pm%]8firS4d 7y\`JnίI R3U~7+׸#m qBiDi*L69mY&iHE=(K&N!V.KeLDĕ{D vEꦚdeNƟe(MN9ߜR6&3(a/DUz<{ˊYȳV)9Z[4^n5!J?Q3eBoCM m<.vpIYfZY_p[=al-Y}Nc͙ŋ4vfavl'SA8|*u{-ߟ0%M07%<ҍPK! ѐ'theme/theme/_rels/themeManager.xml.relsM 0wooӺ&݈Э5 6?$Q ,.aic21h:qm@RN;d`o7gK(M&$R(.1r'JЊT8V"AȻHu}|$b{P8g/]QAsم(#L[PK-![Content_Types].xmlPK-!֧6 +_rels/.relsPK-!kytheme/theme/themeManager.xmlPK-!Ptheme/theme/theme1.xmlPK-! ѐ' theme/theme/_rels/themeManager.xml.relsPK] F 99999<z 6e]< !"#*%&*+k.#01248;V=?%AACYDqFK(*+,./12346789:;=>?ABCDEGHIKLNOPQST;$+2:@DM0Oհ)-05<@FJMRUVY4HJ $ 8 : # 7 9  !Pdfo';=  -AC%'H\^n24Znp02J^`*,o^rt !!K"_"a""""+#?#A#v###$$$W%k%m%%%%%%%&&&(((((()&)())))))),0@0B0]1q1s1t222f4z4|444467J7L77777778$8&8J8^8`8}88888849H9J9~999:::,:@:B:;;;F::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: 25<!!8@2 H(  T h1  #" ?P   "5GP   "w1w  BC DEXF( < x > - r  s ` M0Ix,v @       "hn,wh B S jJ" 5""h  S jJ"ddwh  S jJ""h   S jJ"5^d^4T z>3 #" ?P    "zP    ">3P   " zzP   "\  # #" `zP4P   "z!z!P   "00\  # #" `z!\0@B S  ?   "F  x. tn0t!*wxAA=CCCyCCCC1D9DDDDE0E2EDEHE[E_E`EiElEnEpErEEEEEEEEEFFFFFFFFFFFFFFFFGS  $ #+05 79mo,} V 6!;!%%Y(a((((())&*/***V,_,`,,0041<13333[4b45577+949':,:;;= =7=?=@@AAAAB\B_BjBlB|BBBBBBBBBBCCCC,C.C=CCCXCZChCjCyCCCCCCCCCCCCCCDDDDDD:D@DnDqDDDDDDE0E3EDEIE[E`ElEnEEEEEEEEEEEEEEF6F;F>FCFFFFFFFFFFFFFFFFFFFFF3333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333o'> (n+#B#v##$$W%n%(((())),0C04467M7J8a8}88::,:C:;;<<==q==?AAADDFFFFFFFFFFFFFFFFFFFo'> (n+#B#v##$$W%n%(((())),0C04467M7J8a8}88::,:C:;;<<==q==?AAADDFFFFFFFFFFFFFFFFFFFZYXOc9d C 4 +J B|r<7c` {n$*p.0\2a283m5d7XC9[9?AhC.FJK5-N%P&VR *T]&^'da fYg~lUnonY$pZr;8tsH~XCB|j,K5WCED4#%q8wc!K, r]Vq 3#-FTQZ{"iq%] \xTwV2O bdFF@AAhIAALABEF@&P@J@`@UnknownG* Times New Roman5Symbol3. * ArialA BCambria Math"1hef.jˆ /<$ /<$d4FF2HX ?&^2!xx/Simulation of Measurements of a Random Variable Peter ShermanshermanpCompObjy