Transcript for:
Fundamentals of Introductory Calculus

Okay, welcome to Introductory Calculus. I  will start with some practical information   and then I'll tell you a little bit  about the syllabus and what we will   cover in this course and then give you  some examples of differential equations   from physical sciences and then a little  bit of integration towards the end. So   for many of you this might be the easiest  course here that you'll take in Oxford,   but I think things will get progressively  harder so maybe in a couple of weeks it will   be interesting to everybody if today's lecture  might be a bit too easy for some of you. Okay   so let me tell you some practical information. So  we have 16 lectures. The lecture notes are online. These are the lecture notes which were written  by Cath Wilkins - she taught this course for a   few years until last year so we'll just follow  them I guess. I should have introduced myself   at some point as the lecturer so you can call  me Dan my name is Dan Ciubotaru and we'll meet   twice a week today is special just because  it's the first week. We'll meet on Mondays   and Wednesdays at 10am, so not too early  and you'll have eight problem sheet. So   the first two problem sheets are online.  The eight problem sheets you'll cover in   four tutorials in your college. Okay so  four hours for tutorials. So I said the   lecture notes are online, the reading list  is also online (so see online) the book that   I like is Mary Boas's Mathematical Methods  in Physical Sciences you know this book and most of your colleges should have a copy of  this. If not the university does as well. So   this is this book is quite concise and it has  various examples from physics and engineering   and science it also has the added advantage  that if unlike the other books on on the   reading list if you drop this one on your foot  you might be able to to walk without seeing an   orthopedic surgeon. All right, so that's that 

  • any questions about this? Okay, now syllabus so the first half of the course - about  seven or eight lectures will be devoted   to differential equations. So this is about  seven/eight lectures so of two kinds ordinary   differential equations (ODEs) and partial  differential equations (or PDEs) so I'll   give you some examples very soon. We'll look at  fairly easy examples of differential equations,   we'll learn some techniques. It's a combination  solving them - it's a combination of science and   art. You have to do some educated guesses at  some points but it's quite an interesting and   very useful subject and then after that we'll  talk about line and double integrals and the   reason these are useful is because we will  be able to compute arc lengths of curves and   areas of various regions in the plane or  surfaces. So this is maybe three lectures and then finally we'll do calculus  of functions in two variables. So   this should be viewed as a gentle  introduction into multivariable   calculus. So among the things that we'll  do we'll look at various surfaces gradient normal vectors will look at  Taylor's theorem in two variables,   critical points and a little  bit of Lagrange multipliers, which are useful for optimization problems. Okay,  now there is a lot of interaction between this   course and other premlim courses that you will  take. So intro calculus will be directly useful in   well obviously multivariable calculus as I said.  In a way it's a little bit unfair as we set up the   work and we do some examples for in introduction  calculus but then the really cool result and   theorems you prove in multivariable calculus, so  we just do a little bit of the ground work towards   that. You also do these and they are also useful  in dynamics and in PDEs which you will do next   term fourier series and PDEs. Now there is a lot  of interaction between intro calculus and analysis   particularly analysis two which is what you do  next term so there will be quite a few results   from analysis that will just say and not prove,  maybe prove some particular examples and so on,   but real rigorous proofs you'll do in analysis  next term but then it all comes together when   you revise or for your exams in Trinity. Okay so  there's that. Of course in Part A there will be   lots of applied mathematics options that will  continue this. Differential equations is a big   option, fluid and waves etc so this is a very  useful course it's also mandatory so you have   to be here. So now let me give you some examples  of where all these might appear. Okay so all these DEs, so what is an ordinary differential  equation? So this is an equation involving   an independent variable, let's call it x and a  function of x which we call it y. So y this would   be the dependent variable and the derivatives  of y with respect to x so for example dy/dx   d^2y/dx^2 etc you know so the order of the highest  derivative that occurs we call that the order of   the differential equation. So for example the  simplest so the simplest kind of ODE would be   something of the form dy/dx equals some function  in x so dy/dx equals f(x) - you can solve that   by direct integration so this can be solved  so y equals. So you can think of y as being   the antiderivative of f(x) and then we we can  use integrations that's the simplest kind of   differential equation that we can have and this is  the reason why we'll start the course by reviewing   a little bit of integration techniques. But there  can be more interesting differential equations,   so let me give you some examples from physical  sciences. So for example from mechanics,   this is something that you have all seen,  you can have Newton's second law which   says that the force is the mass times the  acceleration (so a is the acceleration). But then the acceleration is a derivative. It is  the derivative of the velocity with respect to   time. So that's already a differential equation  but it could be a second-order differential   equation if you think that v is dr/dt where r is  the displacement. Then you get for example a is   d^2r/dt^2 which is a second order differential  equation so that that's an easy example of how   differential equations appear in mechanics. Well  you could also have differential equations in   engineering, or if you have an electrical circuit.  So if I take a simple one, so a simple series   circuit, which for example: an RLC circuit -  which means that it has the following components:   it has R stands for resistor, so it has a resistor  R, it has an inductor L with inductance L and it   has a capacitor with capacitance C and it has  a source of voltage something like a battery V.   So I have a capacitor with C capacitance and the  resistor with R resistance and an inductor with L inductance. So here are R, L and C are  constants they're independent of time but then I'm interested for example in the  current across the circuit so this is the   current across I(t) is the current across the  circuit which is a function of the time. So   in terms of differential equations t the time  would be the independent variable and this I(t)   for example is a dependent variable I can  also have Q(t) which is the charge across   capacitor on the capacitor and the relation  between the two of them is that I=dQ/dt so   Kirchoffs law says that the total voltage is  zero around the circuit. Which is in a another   way saying the voltage V from the battery  which is a function of t equals the voltage   across the resistor plus the voltage across the  resistor plus the voltage on the capacitor and   now you write each one of them. The voltage  across the resistor by Ohm's law is R I(t),   the one on the capacitor is just 1/C and for  the inductor is L dI/dt which is Faraday's Law. So now I can express for example so I have an  equation involving V I and Q but I is dQ/dt   so I can rewrite everything in terms of Q, for  example. So I can get a differential equation   in Q which will be simply: this would be the  leading term dI/dt so L times dI/dt becomes L   times d^2Q/dt^2 + RdQ/dt +1/C Q = V, so that  is a second order differential equation that   appears in electrical circuits. Yeah so it's  second order because the highest derivative is   of second order - it has constant coefficients  because the constants R,L and C are constant   and it's what we'll call inhomogeneous because  this doesn't have to be zero. So those are the   type of differential equations that we can we  can study and there are many other examples,   so I'll leave one as an exercise for you. So  I'll tell you the problem is the rate at which   a radioactive substance decays is proportional to  the remaining number of atoms. So I want you to,   as an exercise, to write a differential equation  that describes this situation. Okay so we'll come   back to things like this later. So what  the question is: what's the differential   equation? Okay so as you progress along in this  course in the mathematics course here, you will   encounter very very interesting and sophisticated  differential equations in applied mathematics,   so we're just scratching the surface a little bit.  All right now, going back to what I what I said   before - the simplest kind of ODE is dy/dx equals  f(x) which you can solve by integration so let me   review a couple of facts about integration. So  one of the most useful techniques which I'm sure   most of you are quite familiar with is integration  by parts. Okay so where does integration by parts   come from? Well it comes from the product rule. If  I have two functions f and g and I multiply them   and then I differentiate them so f times g prime  is f prime g plus f g prime, which means that f   g f times g prime equals f times g prime minus  f prime times g and if I integrate both sides then I end up with the integration by parts which  is f times g prime dx if they're functions of   x. It equals f times g minus f prime times g dx.  Okay so this is the version indefinite integral's   version, you can have a definite integral  version where you put the limits of integration. So let me spell it out. So this is the  definite integral's version. Alright,   so let's do a couple of examples. So the first  example - so suppose I want to integrate x^2   sin(x). So this would solve, so this would  give this gives the solution to dy/dx equals   x^2 sin(x) ok so in the integration by parts  you have to decide which one is f and which   one's g. Now clearly I would like to decrease  the power here I know I can never get rid of   the sine by differentiation so then maybe this  then I have to do this f and this is g prime which means that g is -cos x so if I  call this integral I I is x squared   times minus cos x and then minus  the derivative of f which is 2 x   times minus cos x dx. This is minus x  squared cos x plus 2 times x cos x dx. And now again this should be f and this should  be g prime of course x plus 2 times x sine x   minus 2 times sine x dx. So please try to  follow through what I'm doing and let me   know if I make a mistake. This is kind of  my nightmare to integrate integrals like   this while I'm being filmed. Right exactly  what I like to do so 2x sin(x) and then   minus cos(x) then plus C is this... so plus  plus thank you, as I said so C here denotes   a constant because we're doing indefinite  integrals. All right let's do another example. So again an indefinite integral:  2x minus 1 times Ln (x^2+1)dx. Ok. What do you think? Which one should be f and which  one should be g or g prime. Say that again? Right   so this I want to differentiate to get rid of the  logarithm, so I should call this f which means   that this is going to be g prime, thank you and  that makes g: x squared minus x so this becomes   x squared minus x ln (x^2+1) minus the integral  of x squared minus times the derivative of the   natural log of x squared plus 1 which is 2x over  x squared plus 1 dx so and finally this term. What   do I do here? Good so we we do long division, so  let's rewrite it first. This is x squared minus x   ln x squared plus 1 and then minus 2x cubed minus  x squared over x squared plus 1 dx so I have to   remember how to do long division, so I have x  cubed minus x, now depending how you learn this   you will draw the the long division in different  ways so you just do it your way and I'll do it my   way. So that's x minus x cubed minus x then minus  x squared minus 6 and that's minus 1. Ok so this   means that x cubed minus x squared over x squared  plus 1 equals x minus 1 plus minus x plus 1. Did you get the same thing? Okay so let's call this integral J and now we  compute J. The integral of x minus 1 plus minus   x plus 1 over x squared plus 1 dx which equals  x squared minus 1/2 x squared minus 6 and then   how do I integrate this term? The fraction  so I should split x over x squared plus 1 dx. Yeah and let me write the last term plus dx over x  squared plus 1. So this 1, the last term we should   recognize that - what is it? It's arctan(x) or  tan inverse depending how you want to denote it   this is arctan of x which is tan inverse of x.  Now what do we do with this we can substitute. Yeah let's do that. So that we remember how to do  substitutions. You might just look at it and know   what it is right, but just to review substitution  if I said u equals x squared plus 1 then du equals   2x dx, so du/dx equals 2x which means that this  is 1/2 d u / u which is 1/2 Ln of u which is 1/2   Ln of x squared plus 1 that you might have guessed  just because you have enough practice some of you.   Okay so now let's put them all together.  So J is 1/2 x squared minus x minus 1/2   Ln x squared plus 1 plus tan inverse of x + some  constant, which means that the original integral,   the integral in the beginning which I should  have called I so that I don't have to roll   down the boards, that equals x squared minus x  Ln x squared plus 1 minus twice this. So minus   x squared plus 2x plus Ln x squared plus  1 minus tan inverse x and then plus 2c. Thank you. Any other mistakes? Alright. Okay so that's a that's an intro. There are cases  when integration by parts will not simplify either   of the two functions f and g but what happens is  if you do it twice then you sort of come back to   what you started with. So the typical example is  I equals the integral e to the x times sine x dx.   So maybe we don't need to go through the entire  calculation - this is in the lecture notes as   well - but how would you solve it? Right so  you do it so for example I can say that this   is g prime and this is f and then I integrate  I get cos and then I do it again and I will end   up with some expression - they seem to grow  and then I solve for it. So you do this and   you get the answer to be something like 1/2 e  to the x sin(x) minus cos x then plus constant. Okay so another type of example which  are more difficult are the ones which   you cannot solve in just one go  but you have to find a recursive   formula. So I'll just do an example like  that. You've seen other examples before   so this is when we get a reduction or if  you want to call it a recursive formula. So I start, suppose I'm looking at this integral:  cosine to the n x dx. Now I want to label this   integral I(n) because I'm going to get a formula  of I(n) in terms of I(n-1) or I(n-2) etc. Now   there is not much choice here what you should  call f and what you should call g so I'm going   to just do it. So this is cos^(n-1) x times cos  x dx so this is f and this is g prime, then we   get cos n minus 1 x sin(x) minus the integral.  Now I need to differentiate f so (n-1) cos n   minus 2 x and then minus sine x and then another  sine x dx, which equals cos n minus 1 x sine x. - n - 1 times, or maybe I'll make it a plus, and  minus 2 x sine squared x dx. So if I write it like   that what do you do now? You write sine squared as  1 minus cos squared x, which then gives you cos n   minus 1 x sine x plus n minus 1, the integral of  cos n minus 2 x dx minus n minus 1 the integral of   cos n squared x dxm so now I recognize that this  is the integral of course n minus 2 is I sub n   minus 2 and the integral of cos, and this is I(n)  so I have I(n) equals that so if I solve for I(n)   we get I(n) that n I(n) equals cos^(n-1)(x)sin(x)+  (n-1)I(n-2), which gives me the recursive formula I(n) equals 1 over n cos (n minus 1) x  sine x plus n minus 1 over n. So this   is true for all n greater than or equal  to 2. Okay now if I want to know all of   these integrals I(n) using this formula 
  • what else do I need to know? I(0) and   I(1) because it drops down by 2. So let's  compute I(0) and I(1), so I(0) would be just   the integral dx which is x plus C, and I(1)  is the integral of cos x dx which is sine x plus C and now with this you can  you can get any integral you want,   for example if you want to get I don't know I(6), you just follow that and you get that  it's 1/6 cos^5(x)sin(x) plus 5 over   6 times I(4) which is 1 over 6 cos^5(x) sine x  plus 5 over 6 times I(4) is 1/4 cos cubed x sine x   plus 3/4 I(2), then what is I(2) is 1/2 cos x sine  x plus one-half I zero but I zero is x so you put   your substitute this in there and I get an I six  is 1/6 cos to the fifth x sine x plus five over 24   cos cubed x sine x plus 5 times 3 times 1 over 6  times 4 times 2 cos x sine x plus 5 times 3 over over 6 times 4 times 2 x, so it has I think  you can you can probably cook up a general   formula using this example - you see how  it goes. So if I asked you to write a sum   involving all the terms I think you can you can  get the coefficients of each term inductively. Okay so this is a quick review of integration  by parts. If you're not fully comfortable with   these examples or similar examples, then  get get an integration textbook and do a   few more examples with integration by parts,  substitutions and so on because in solving   differential equations we'll learn a lot of  techniques but ultimately you will have to   integrate some functions so you should be able  to do that. What we learn is how to reduce the   problem to integrating various functions  but you'll have to be able to do that. Okay so we discussed about  the simplest kind of these   which can be solved just by direct  integration. The next simplest of these are the so-called separable. so we had the case dy/dx equals f(x)  which you can just integrate the next   case would be dy/dx equals a(x) b(y).  So what I mean by that is that this is   a function in only x and similar  b of y is a function of y only. If you have a situation like that then you can  reduce it to the direct integration with one   simple trick. If b(y) is not zero then you divide  by it and you get 1 over b(y) dy dx equals a(x), and now you can integrate just as we did  before. So you'll get then the integral so the left-hand side is the integral dy  over b(y) and the right-hand side is a(x)dx   and now you have to direct integrations  which hopefully we can we can solve,   right. The type of integrals that we have  in this course will be the kind for which   you can apply integration by parts or some  other techniques and solve them if I if I   were to write an arbitrary function there and  ask you how to integrate it then we can't do   that in a closed formula. Okay so here's  an example: find the general solution to the separable differential equation. So the  hint is already in the problem that this is   a separable differential equation x times  y squared minus 1 plus y x squared minus   1 dy/dx equals to 0 and x is between 0 & 1 to  avoid some issues about continuity or whatnot. Okay, how do you separate this differential  equation? How do you separate the variables?   Correct, so I think you're about two steps  ahead of me. It's correct but let me do it   step by step. So what I will do is first  isolate that so I have y x squared minus   one dy/dx equals minus xy squared minus one  and then separate the variables, as the name   suggests. You have y over y squared minus one  dy/dx equals minus x over x squared minus one. Okay what do we do now? Correct, so if we look  at this then, so we integrate, let's integrate,   well let me write one more, so we integrate this  and we get y over y squared minus one dy equals   minus x over x squared minus 1 dx so now we could  do substitution as we did before but I think we   know how to do it, this looks like the derivative  of a logarithm. So if I differentiate ln of x   squared minus 1 then I get 2x over x squared minus  1, so except x is between 0 & 1 so maybe it's   better to write this as 1 x over 1 minus x squared  and get get rid of the minus sign. So then I'll do   1 minus x squared minus 2x over 1 minus x squared  so then this is minus Ln of 1 minus x squared and a half and then plus a C. Whereas  here I will have to put absolute values   because I don't know. It's 1/2 Ln so y  squared minus 1 in absolute values right. Now the easiest way to write this is to get rid  of the logarithm by moving this to the other side.   Using the properties of the logarithmic, so let's  do that. So I have 1/2. If I move the logarithm in   x to the left hand side then I use the property,  well it doesn't matter much, it equals C which   means that the equation will be y squared minus  1 times 1 minus x squared absolute value equals   it would be e to C squared or e to to C which I  can just call another kind of C and this would   be a positive number so the equation then  that we get is so the the answer then is where C is positive. But I can relax that, this  is always positive because 1 minus x squared   is always positive, because I'm assuming x is  between 0 & 1 but I can rewrite the answer in   a nicer form by dropping the absolute value and  dropping the assumption on C. So another way of   this for uniformity, I'll write it as 1 minus  y squared equals C so 1 minus y squared times   1 minus x squared equals C. No assumption on C  except so C could be both positive or negative   except in this formula it looks like it can't  be 0, right? This here I got an exponential   which is never 0 so this is positive, I drop the  absolute value and now C can also be negative but   somehow 0 is missing. How is that possible?  That doesn't look like solid mathematics? Yes? That's right. Okay, so where did I lose  that case? Right here. So I divided by   y squared minus one. I did that ignoring  the case when y squared minus one is zero,   so note so let's call this star here so in  star y -1 is y squared minus 1 is not 0 but   if we need to allow that because it is possible  for y to be plus or minus 1 for example. If y   is the constant function 1 then this is 0  and dy/dx is 0, so that's okay. So if we   allow if y is plus minus 1 is included in the  solution if we allow C to be 0 in the answer. So then the bottom line is that the answer is   this implicit equation in y and x  where C can be any constant. Good. So be careful when when you divide by  the function in y, as I said here you   can do that if you know that's not zero,  but sometimes you get solutions from it   being zero so you have to be careful there. All  right, that's the end of the first lecture I'll   see you tomorrow for for the second lecture  and we'll do more differential equations!