Transcript

the following content is provided under a Creative Commons license your support will help MIT open courseware continue to offer highquality educational resources for free to make a donation or to view additional materials from hundreds of MIT courses visit MIT open courseware at ocw.mit.edu to begin today I want to remind you I need to write it down on the board at least twice of the fundamental theorem of calculus we called it ftc1 because it's the first version of the fundamental theorem we'll be talking about another version called the the second version today and what it says is this if F Prime is equal to F then the integral from A to B of f ofx DX is equal to F of B minus F of a so that's the fundamental theorem of calculus and the way we used it last time was this was used to evaluate integrals not surprisingly that's how we used it but today today I want to reverse that point of view we're going to read the equation backwards and we're going to write it this way and we're going to use f to understand capital f in other words the the the derivative to understand the function so that's the reversal of point of view that I'd like to to make and we'll we'll make this point in in in various ways so information about F about fime uh gives us information about F now since there were questions about the mean value theorem I'm going to illustrate this first by making a comparison between the uh fundamental theorem of calculus and the mean value theorem so we're going to compare this fundamental theorem of calculus with what we call the mean value theorem and in order to do that I'm going to introduce a couple of notations I'll write Delta f is f of B minus F of a and another highly imaginative notation Delta X is equal to B minus a so here's the change in F there's the change in X and then this fundamental theorem can be written of course that's right up above there as the form formula and it's the formula for Delta F so this is what we call the fundamental theorem of calculus I'm going to divide by Delta X now and if I divide by Delta X that's the same thing as 1 over B minus a times the integral from A to B of f ofx DX all right so I've just Rewritten the formula here and this expression here on the right hand side uh is a fairly important one this is the average of f that's the average value of f now now so this is going to permit me to make the the comparison between the mean value the which we don't have stated yet here and the uh fundamental theorem and I'll do it in the form of inequalities so right in the middle here I'm going to put the fundamental theorem it says that Delta F in this notation is equal to well if I multiply by Delta X again I can write it as the average of f so I'm going to write it as the average of f Prime here times Delta X all right so we have this Factor here which is the average of f prime or the average of little F it's the same thing and then I multiplied through again so I put the thing in the in the red box right here isn't what the average of big F so the question is why is this the average of little F rather than the average of big f um so the average of a function is the typical value if for example little F were a constant okay little F were a constant then this integral would be uh so the the question is why is this the average and I'll I'll take a little second to explain that but I think I'll I'll explain it over here because I'm going to erase it so the idea of an average is the following for example imagine that a is equal to zero and B is equal to n let's say for example and uh so we might average we might sum the function from 1 to n all right now that would be the sum of the values from one to n but the average is we divide by n here all right so this is the average and this is a kind of remon sum representing the integral from 0 to n of f ofx DX where the increment Delta X is one all right so this is the notion of an average value here but in the Continuum setting as opposed to the discrete setting whereas uh what's on the left hand side is the change in F the capital F and this is the average of the little F okay so an average is a sum and it's like an integral all right so so in other words what I have here is that the change in F is the average of its infinitesimal change times the amount of uh time elapsed if you like so this is the the the statement of the fundamental theorem just Rewritten exactly what I wrote there but I multiplied back by Delta X now let me compare this with the mean value theorem the mean value theorem also is an equation the mean value theorem says that this is equal to frime of C * Delta X now I I pulled a fast one on you I I used Capital FS here to make the analogy clear but the the role of the letter uh is is is important to make the transition to this comparison we're talking about the function capital F here and its derivative now this is true so now I claim that this thing is fairly specific whereas this unfortunately is a little bit vague and the reason why it's vague is that c is just somewhere in the interval so Su C sorry this is Su C in between A and B so really we don't since we don't know where this thing is we don't know which of the values it is we can't say what it is all we can do is is say well for sure it's less than the largest value say the maximum of f Prime Time Delta X and the only thing we can say for sure on the other end is that it's less than or equal uh to sorry it's greater than or equal to the minimum of frime * Delta X over that same interval this is over 0 less than sorry a less than x less than B all right so that means that the fundamental theorem of calculus is a much more specific thing and indeed it gives the same conclusion it's much stronger than the mean value theorem it's way better than the mean value theorem in fact as soon as we have integrals we can abandon the mean value theorem we don't want it it's too simpleminded and what we have is something much more sophisticated which we can use which is this so it's obvious that if this is the average the average is less than the maximum so it's obvious that it works just as well to draw this conclusion and similarly over here with the with the minimum okay the average is always bigger than the minimum and smaller than the maximum all right so this is the uh connection if you like and I'm going to going to elaborate just one step further by uh talking about the uh problem that you had on the exam so there was an exam 2 problem and I'll show you how it works uh uh using the mean value theorem and how it works using integrals but I'm going to have to use this notation capital F so capital F Prime as opposed to little F which was what was the notation that was on your exam so we had this situation here this these were the Givens of the problem and then the question was um uh the mean value theorem says or implies if you like it doesn't say it but it implies it sorry implies um a is less than F uh sorry capital F of 4 is less than b for which A and B so let's take a look at what the uh what it says well um the mean value theorem says that F of 4 minus f of0 is equal to frime of C * 4 - 0 this is this fime * Delta X this is the the change in X and that's the same thing as 1 over 1 + C * 4 and so the range of values of this number here is from 1 over 1 + 0 * 4 that's four to that's the largest value to the smallest that it gets which is 1 over 1 plus 4 * 4 right that's that's the range and so the conclusion is is that uh F of 4 minus F of 0 is between well let's see it's between four and 4 fths right are those two numbers down there and if you remember that F of 0 is one this is the same as F of four is between 5 and 95s all right so that's the way that uh you were supposed to solve the problem on the on the exam on the other hand let's compare to what you would do with the uh fundamental theorem of calculus with the fundamental theorem of calculus we have the following Formula F of 4 minus F of 0 is equal to the integral from 0 to 4 of DX over 1 + x that's what the fundamental theorem says and now I claim that we can get these same types of results by very Elementary observation it's really the same observation that I made up here that the average is less than or equal to the maximum which is that the biggest this can ever be is uh let's see the biggest it is is when X is is zero that's one so the biggest it ever gets is this and that's equal to four all right on the other hand the smallest it ever gets to be it's equal to this the smallest it ever gets to be is the integral from 0 to 4 of a fifth DX because that's the lowest value that the integrant takes right when X is equal to 4 it's a fifth and that's equal to four fifths okay now there's a little tiny detail which is that really we know that this is the area of some some rectangle and this is strictly smaller and we know that this is these inequalities are actually strict but uh that's a minor point and certainly not one that we'll pay close attention to but now let me let me show you what this looks like geometrically so geometrically we interpret this as the area under a curve here's the uh a piece of a of the the curve uh uh y equals 1 over 1 + x and it's going up to four and starting at zero here and the first estimate that we made that is the upper bound was by trapping this in this in this big rectangle here right we compared it to the constant function which was one all the way across this is yal one and then we also trapped it from underneath by the function which was at the bottom and this was y = 1 15 and so what this really is is these things are the the simplest possible remon sum sort of a silly remon sum this is a remon sum with one rectangle this is the simplest possible one and so this is a very very crude estimate and you can see it misses by a mile the the larger and the smaller values are off by a factor of five but um this one is called the this one is the lower remon sum and that one is less than our actual integral Which is less than the upper remon sum okay and you should by now have looked at those upper and lower sums on your homework so it's just the rectangles underneath and the rectangles on top all right so at this point we can literally abandon the mean value theem because we have a much better way of getting at things if we chop things up into more rectangles we'll get much better numerical approximations and if we use simple-minded expressions of with integrals we'll be able to figure out any bound we could get using the mean value theorem so that's not the relevance of the mean value theorem I'll I'll explain to you why we talked about it even uh in a few minutes okay are there any questions before we go on yeah did uh I knew that the range of C was from 0 to four I should have said that right here this is true for this theorem the mean value theorem comes with an extra statement which I missed which is that this is for some C between 0 and four so I know the range is between 0 and four the reason why it's between 0 and four is that's part of the mean value theorem we started at zero we ended at four so the C has to be somewhere in between that's part of the mean value theorem that that's the question is do you exclude any values that are above four and below zero yes absolutely the point is that in order to figure out how F changes capital F changes between 0 and four you need only pay attention to the values in between you don't have to pay any attention to what the function is doing below zero or above four there those things are strictly irrelevant yeah yeah I mean it's it's strictly in between these two numbers I have to understand what the lowest and the highest one is that's it's approaching that so all right okay so now uh the next thing that we're going to talk about is since I've got that one up there the fundamental theorem of calculus one I need to talk about uh version two so here's the fundamental theorem of calculus version two I'm going to start out with a function little F and I'm going to assume that it's continuous and then I'm going to define a new function which is defined as a definite integral G of X is the integral from a to X of f of T DT now I I want to emphasize here because it's the first time that I'm writing something like this that this is a fairly complicated Gadget it plays a very basic but and very fundamental but simple role but it nevertheless is a little complicated What's Happening Here is that the upper limit I've now called called X and the variable T is ranging between a and x and the A and the X are fixed when I calculate the integral and the t is the What's called the dummy variable it's the variable of integration you'll see a lot of people who will mix this x with this T and if you do that you will get confused potentially hopelessly confused in this class in 1902 you will be completely lost if you do that so don't do it don't mix these two guys up all right it's actually done by many people in textbooks and it's fairly careless especially in old-fashioned textbooks but don't do it all right so here we have this G of X now remember this G ofx really does make sense if you give me an A and you give me an X I can figure out what this is because I can figure out the remon sums of course I need to know what the function is too but anyway we have a numerical procedure for figuring out what the function G is now as is suggested by this mysterious letter X being in the place where it is I'm actually going to vary X so the conclusion is that if this is true and this is just a a parenthesis not part of the theorem it's just an indication of what the notation means uh then G Prime equals F let me first explain what the significance of this theorem is from the point of view of differential equations G of X solves the differential equation y Prime equal F ofx so y Prime equals f i shouldn't you put the X in if I got it here with the condition y of a is equal to zero so it solves this pair of conditions here the rate of change and the initial position is is specified here because when you integrate from a to a you get zero always and what this theorem says is you can always solve that equation when we did differential equations I said that already I said we'll treat these as always solved well here's the reason we have a numerical procedure for computing things like this we can always solve this equation and the formula is a fairly complicated Gadget but uh so far just associated with remon's sums all right all right now uh let's just do one example and then we'll let's just do one one example unfortunately not a not a complicated example and maybe not persuasive as to why you would care about this just yet but nevertheless very important because this is the quiz question which everybody gets wrong until they practice it all right so the integral from say 1 to X of d t/ t ^2 let's try this one here so here's an example of this theorem I claim now this is a question which challenges your ability to understand what the question means because it's got a lot of symbols it's got the integration and it's got the differentiation however what it really is is an exercise in recopying you look at it and you write down the answer and the reason is that by definition this function in here is a function of of the form G ofx of the of the theorem over here and by defin so this is the G ofx and by definition we said that g Prime of X is equal to F ofx well what's the f ofx look inside here it's What's called the integrand this is the integral from 0 to X of f of T DT right where the F of T is equal to 1 over t^2 so your ability is challenged you have to take that 1 over t^2 and you have to plug in the letter X instead of t for it and then write it down as I say this is an exercise in recopying what's there right so this is quite easy to do right I mean you just you just look and you write it down but nevertheless it looks rather looks like a long elaborate object here pardon me so the question was why did I integrate why did I not integrate ah very good question why did I not integrate the reason why I didn't integrate is I didn't need to just as when you take the anti-derivative sorry the derivative of something you take the anti-derivative you get back to the thing so in this case we're taking the anti-derivative something and we're differentiating it so we end back in the same place where we started started we started with f of T we're ending with f little F all right so you integrate and then differentiate and you get back to the same place now the only difference between this and the other version is in this case when you differentiate and integrate you could be off by a constant that's with that shift why there two pieces to this one but there's never an extra piece here there's no plus C here when you integrate and differentiate you kill whatever the constant is because the derivative of a constant is zero all right so no matter what the constant is hiding inside of G you're getting you're getting the same result all right so this is this is the the uh basic idea now I just want to double check it using the fundamental theorem of calculus one here so let's actually evaluate the integral so now I'm going to do what you suggested which is I'm just going to check whether it's true no no I am I'm going to just double check that it's consistent and certainly is is slower this way and we're not going to want to do this all the time but we might as well check one so this is this is our integral and we know how to do it no I need to do it and this is minus tus one evaluated at 1 and X again there's something subliminally here for you to think about which is that remember its T is ranging between one and and T equal x and this is one of the big reasons why this letter T has to be different from X because here it's one and there it's X it's not X so you can't put an X here again this is tal 1 and this is tals X over there and now if I plug that in I get what uh I get minus one /x and then I get minus a minus1 all right so this is let me get rid of those little T's there so this is a little easier to read all right and so let now let's check it it's D by DX so here's what G of X is G of X is 1 - 1 /x that's what G of X is and if I differentiate that I get plus 1x^2 that's it you see the constant washed away now so now here's my job my job is to prove these theorems I never did prove them for you so I'm going to prove the fundamental theorem of calculus but I'm going to do two first and then I'm going to do one and it's just going to take me uh just one Blackboard it's not that hard the proof is by picture and using the interpretation as area under the curve so if here's if if here's the value of a and this is the graph of the function uh yal F ofx then I want to draw three vertical lines one of them is going to be at X and one of them is going to be at X Plus Delta X so here I have the interval from 0 to X and next I have the interval from X to x uh Delta X more than that and now the pieces that I've got are the area of this part so this has area which has a name it's called G of X by definition G X which is sitting right over here in the fundamental theorem is the integral from a to X of this function so it's the area under the curve so that area is G of X now this other chunk here I claim that this is Delta G this is the change in G it's the value of G of X that is the area of the whole business all the way up to X Plus Delta X minus the first part G of X so it's what's left over it's the it's the incremental amount of area there and now I just am going to carry out a a a pretty standard estimation here this is practically a rectangle and it's got a base of Delta X and so we need to figure out what its height is this is Delta G and it's approximately its base times its height but what is the height well the height is maybe either this segment or this segment or something in between but they're all about the same so I'm just going to put in the value at the first point that's the Left End there so that's this height here is f ofx so this is f ofx and so really I appro imated by that rectangle there and now if I divide and take the limit as Delta X goes to zero of Delta G over Delta X it's going to equal F ofx and this is where I'm using the fact that f is continuous because I need the values nearby to be similar to the value and the limit okay that's the end this is the this is the the end of the proof so I'll put a nice little QED here all right all right so we've done fundamental theorem of calculus 2 and now we're ready for fundamental theorem of calculus one all right so now uh I still have it on the Blackboard to remind you it says that the integral of the derivative is the function at least the difference between the values of the function at two places so the place where we start is that with this with this property that uh fime is equal to F that's the starting that's the hypothesis now unfortunately I'm going to have to assume something extra in order to use uh the fundamental theorem of calculus 2 which is I'm going to assume that f is continuous that's not really necessary but that's just a very minor technical point which I'm just going to ignore all right so we're going to start with f Prime being equal to little F and then I'm going to go somewhere else I'm going to define a new function G ofx which is the integral from a to X of f of T DT now this is where we needed all of the labor of reman sums because otherwise we don't have a way of making sense out of what this even means all right so hiding behind this one sentence is the fact that we actually have a number we have a formula for such functions so there is a function G ofx which once you've produced a little f for me I can cook up a function capital G for you now we're going to apply this fundamental theorem of calculus 2 the one that we've already checked so what does it say it says that g Prime is equal to F and so now we're in the following situation we know that fpre of X is equal to G Prime of x all right that's that's what we've got so far and now we have one last step to get a good connection between f and g which is that we can conclude that f ofx is G of X plus a constant now this little step may seem innocuous but I remind you that this is the spot that requires the mean value theorem so in order not to lie to you we actually tell you what the underpinnings of all of calculus are and they're this the fact if you like that if two functions have the same derivative they differ by a constant or that if a function has derivative zero it's uh a constant itself that's that is the the fundamental step that's needed the underlying step that's needed and unfortunately there aren't any proofs of it that are less complicated than using the mean value theorem and so that's why we talk a little bit about the mean value theorem because we don't want to lie to you about what's really going on yes the question is how did I get from here to here and the answer is that if G Prime is little F and and we also know that frime is little F then frime is G Prime okay other questions all right so we're almost done I just have to work out the arithmetic here uh so I start with f of B minus F of a and that's equal to G of b + C minus G of a plus c and then I cancel the C's so I have here G of B minus G of a and now I just have to check what each of these is so remember the definition of g here G of B is just what we want the integral from A to B of f ofx DX well I call it f of tdt that's the same as F ofx DX now because I have the limit being B and I'm allowed to use x as the dummy variable okay now the other one I claim is zero because it's the integral from a to a all right this one is the integral from a to a which gives us zero so this is just this minus zero and that's the end that's it this is I started with f of B minus F of a I got to the integral question how did I get from F of B minus F of a is G of B plus C minus G of a plus c that's the question there oh sorry this is an equal sign sorry the second line didn't draw okay equals because we're plugging in for f ofx the formula for yes you this this step here or this one the um right so that was a that was a good question but the the the the answer is that that's the statement that we're aiming for that's the fundamental theorem of calculus one which we don't know yet so we're trying to prove it and that's why we haven't we can't assume it yeah okay so so so let me just notice that in the example that we had before we go on to something else here in the example above what we had was was the following thing we had uh we had say f ofx is equal to -1x X so frime of X is 1x^2 and say G of X is equal to 1 - 1 /x and you can see that either way you do that if you integrate from 1 to two let's say which is what we had over there DT over t^ squared you're going to get either um - 1 / T 1: 2 or if you like 1 - 1/t 1:2 so this is the F version this is the the G version and that's what plays itself out here in this in this uh General proof all right so now I want to go back to the theme for today which is using little little F little F to understand uh capital f in other words using F the derivative of f to understand capital F and I want to illustrate it by some more complicated examples so I guess I just erased it but um we just took the an derivative of 1/ t^2 and there's all of the powers work easily but uh one and the tricky one is the power 1 /x so let's consider the differential equation L Prime of X is equal to 1 /x and say with the initial value L of 1 is equal to zero this the solution so the fundamental theorem of calculus 2 tells us the solution is this function here L of x equals the integral from 1 to X DT / T that's how we solve all such equations we just integrate take the definite integral and I'm starting at one because I insisted that L of one be zero so that's the solution to the problem and now the thing that's interesting here is that we started from a polinomial or we started from a rational a ratio of polinomial that is 1 over t or 1/x and we get to a function which is actually what's known as a transcendental function it's not an algebraic function Yeah question over here we say1 How uh the question is why is this equal to that right and the answer is it's for the same reason that this is equal to that it's the same reason as this it's that the ones cancel you've taking the value of something at two minus the value at one the value at two minus the value at one and you'll get a one in the in the one case and you get a one in the other case and then you subtract them and they will cancel it'll give you zero these two things really are equal this is not a function evaluated at one place it's the difference between the function evaluated at two and the value at one and whenever you subtract two things like that constants drop out that's right if I put two here if I put C here it would have been the same it would just have dropped out it's not it's not there and that's exactly this arithmetic right here it doesn't matter which anti-derivative you take when you take the differences the C's will cancel you'll always get the same answer in the end that's exactly why I wrote this down so that you would see that it doesn't matter which one you do all right so we still have a couple of minutes left here um this is actually so let me go back so here's the anti-derivative of 1 /x with value one at zero now in Disguise we know what this function is we know this function is the logarithm function but this is actually a better way of deriving all of the uh formulas for the logarithm this is a much quicker and more efficient way of doing it we had to do it by very laborious processes this will allow us to do it very uh easily uh and so uh I'm going to do that next time uh but rather than do that now I'm going to point out to you that we can also get truly new functions okay so there are all kinds of new functions so the first example of this time kind would be for example to solve the equation Y Prime is equal to eus x^2 with y 0 is equal to0 let's say now the solution to that is aun function which again I can write down by the fundamental theorem it's the integral from 0 to X of eus T ^2 DT this is a very famous function this shape Here is known as the bell curve and it's the thing that comes up in probability all the time this shape eus x^2 and our function is geometrically just the area under the curve here this is f ofx if this places X so I have a geometric definition I have a way of constructing What It Is by reman sums and I have a function here but the Curious Thing about f ofx is that F ofx cannot be expressed in terms of any function you've seen previously so uh logs exponentials trig functions cannot be it's a totally new function nevertheless we'll be able to get any possible piece of information we would want to out of this function it's perfectly acceptable function it will work just great for us just like any other function just like the log and what this is analogous to is the following kind of thing if you take the circle the ancient Greeks if you like already understood that if you have a circle of radius one then its area is pi so that's a geometric construction of what you could call a new number which is outside of the realm of what you might expect and the weird thing about this number Pi is that it is not the root of an algebraic equation with rational coefficients it's what's called transcendental meaning it's just completely outside of the realm of algebra and indeed the logarithm function is called a transcendental function because it's completely out of the realm of algebra it's only in calculus that you come up with with this kind of thing so so these kinds of functions will have access to a huge class of of new functions here all of which are important Tools in science and engineering so see you next time