Solving A Differential Equation | Two Methods

แชร์
ฝัง
  • เผยแพร่เมื่อ 17 ม.ค. 2025

ความคิดเห็น • 56

  • @duaafatima-nk3bo
    @duaafatima-nk3bo ปีที่แล้ว +1

    nice contribution for this era people

  • @IantoCannon
    @IantoCannon ปีที่แล้ว +5

    That first substitution was filthy

  • @zahari20
    @zahari20 ปีที่แล้ว +22

    The best substitution for the integral is x = sinh(t). Then you get integral of cosh(t)^2 where cosh(t)^2 = (1/2)(1+cosh(2t)) etc.

    • @adandap
      @adandap ปีที่แล้ว +4

      Yes, for sure. It falls out in half a dozen lines when you do that.

    • @marciliocarneiro
      @marciliocarneiro ปีที่แล้ว +1

      really.. you´re right..

  • @Shadi677
    @Shadi677 ปีที่แล้ว +6

    The tan approach is amazing .. so we can replace the "ln" at the end with arc sinh(x)

  • @ismaelcastillo188
    @ismaelcastillo188 ปีที่แล้ว +3

    In this kind of integrals is usually clever to use t=sinhx as cosh^2-sinh^2=1

  • @marciliocarneiro
    @marciliocarneiro ปีที่แล้ว +2

    I used the transformation x=tan(u) and obtained a shorter resolution because I obtained the integral of sec(u)^3 which is less complicated

  • @kingoreo7050
    @kingoreo7050 ปีที่แล้ว +4

    Thank you for making more calculus and trigonometry videos!

  • @RaviBRUH
    @RaviBRUH ปีที่แล้ว +2

    you can directly use byparts in the question

  • @Mephisto707
    @Mephisto707 ปีที่แล้ว +4

    10:30 Isn't the square root of (sec x)^2 absolute value of (sec x)?

    • @usdescartes
      @usdescartes ปีที่แล้ว

      It WOULD be, but what we are really doing is using the definitions via the sides of a triangle... which, as you point out, are always positive. But don't forget that x is inside a square root (which is always positive itself), and when we do a back-substitution it still will be. It (almost) never happens that a situation arises where it would matter, so just like if our final answer had a ln|x| in it but we knew x was always positive, we'd write it ln x and get rid of the absolute value bars, we ignore the |sec x| and just cal it sec x because we MUST be in a region where that square root will always be positive anyway.
      I imagine it DOES get tricky if we are doing complex analysis, but my last complex analysis class was 30 years ago, and I don't remember if/how we handled trig sub for those problems... if we even did any.

  • @phild8095
    @phild8095 ปีที่แล้ว +3

    Did all this 45 years ago, worked in process engineering and never used it. Advanced Stats would have been more of a help.

    • @SyberMath
      @SyberMath  ปีที่แล้ว +2

      I don't like stats. No offense

  • @yoav613
    @yoav613 ปีที่แล้ว +2

    Very nice! It can be solved also by IBP (x)' sqrt(1+x^2). Another nice substitution for this integral x=i sinu and you can use the identity i arcsin(-ix)=arcsinhx.

  • @ΚαυλιΜαυρο
    @ΚαυλιΜαυρο ปีที่แล้ว

    Great video!
    More DEs would be great

  • @93techie
    @93techie ปีที่แล้ว +4

    When you replace ln|t|, why did you use t = sqrt(x^2+1)+x? shouldn't it be sqrt(x^2+1)-x?

    • @adamfiser7645
      @adamfiser7645 ปีที่แล้ว

      Yes, there (9:10) should be either +ln(sqrt(x^2+1)+x) or -ln(sqrt(x^2+1)-x), because t^(-1) = sqrt(x^2+1)+x. He confused t and t^(-1).

    • @f1sk313
      @f1sk313 ปีที่แล้ว

      Там действительно ошибка

  • @krunoslavregvar477
    @krunoslavregvar477 ปีที่แล้ว +1

    Nice as well as educational.
    👍✌️👍

  • @joyli9893
    @joyli9893 ปีที่แล้ว +1

    WOW! Integral dy is really equal to y!

  • @behrensf84
    @behrensf84 ปีที่แล้ว +1

    Darn… it’s been so long I don’t remember how to do any of these… I just use a spreadsheet now…

  • @del66404
    @del66404 ปีที่แล้ว +1

    👏👏👏👏👏👏👏

  • @jimmykitty
    @jimmykitty ปีที่แล้ว +2

    Awesome ♥️

    • @SyberMath
      @SyberMath  ปีที่แล้ว +1

      Thanks 🤗

    • @SyberMath
      @SyberMath  ปีที่แล้ว +1

      Long time, no see! 😊

  • @manjoker
    @manjoker ปีที่แล้ว +2

    please solve this is integral ln(cscx+cscx cotx) dx

    • @fartsniffa8043
      @fartsniffa8043 ปีที่แล้ว +1

      its not an elementary integral.

  • @duckymomo7935
    @duckymomo7935 ปีที่แล้ว

    Trig substitution is the best choice here

    • @chadmongoose
      @chadmongoose 5 หลายเดือนก่อน

      Can't we directly use integration by parts

  • @giuseppemalaguti435
    @giuseppemalaguti435 ปีที่แล้ว +1

    y=(1/2)(xsqrt(1+x^2)+arcshx)+c

  • @shafin3365
    @shafin3365 ปีที่แล้ว +3

    👍

  • @angelmendez-rivera351
    @angelmendez-rivera351 ปีที่แล้ว +1

    Let y : U -> R such that y is continuously differentiable everywhere, and such that D(y)(x)^2 - 1 = x^2 everywhere. This is equivalent to D(y)(x) = (-1)^m•sqrt(x^2 + 1) everywhere, with m = 0 or m = 1. Let g[m](x) = (-1)^m•sqrt(x^2 + 1) everywhere. Now, D(y)(x) = g[m](x) everywhere, with m = 0 or m = 1. Here, we can apply the Riemann integral on [0, t] to both sides of the equation, and by the fundamental theorem of calculus, the Riemann integral of D(y) is equal to y(t) - y(0). What remains is for us to find the Riemann integral g[m].
    Notice that sqrt(x^2 + 1) = sqrt(sinh(arsinh(x))^2 + 1) = sqrt(cosh(arsinh(x))^2) = |cosh(arsinh(x))| = cosh(arsinh(x)) = cosh(arsinh(x))^2•1/sqrt(x^2 + 1) everywhere. This is important, because if h(x) = 1/sqrt(x^2 + 1) everywhere, then D(arsinh) = h. Also, notice that if j(x) = x/2 + sinh(2•x)/4 = x/2 + sinh(x)•cosh(x)/2 everywhere, then D(j)(x) = 1/2 + cosh(x)^2/2 + sinh(x)^2/2 = 1/2 + cosh(x)^2 - cosh(x)^2/2 + sinh(x)^2/2 = 1/2 + cosh(x)^2 - 1/2 = cosh(x)^2. As such, we can write that g[m] = (-1)^m•(D(j)°arsinh)•D(arsinh) = (-1)^m•D(j°arsinh) = D((-1)^m•(j°arsinh)), so the Riemann integral must be equal to (-1)^m•j(arsinh(t)) - (-1)^m•j(arsinh(0)). Thus, we have y(t) - y(0) = (-1)^m•(j(arsinh(t)) - j(arsinh(0))) = (-1)^m•j(arsinh(t)) everywhere. j(arsinh(t)) = arsinh(t)/2 + t/2•sqrt(t^2 + 1), hence y(t) - y(0) = (-1)^m/2•(arsinh(t) + t•sqrt(t^2 + 1)) everywhere.

  • @imonkalyanbarua
    @imonkalyanbarua ปีที่แล้ว +3

    Fantastic solution! I was just wondering, what if we bring x^2 to the LHS and take 1 to the RHS. Then we get a form:
    a^2 - b^2 = 1
    Can't we then solve it like a Diophantine equation? It will be easy although the answers will be different of course, but is it wrong? Kindly explain it to me. Thank you.

    • @gdtargetvn2418
      @gdtargetvn2418 ปีที่แล้ว +1

      what the

    • @imonkalyanbarua
      @imonkalyanbarua ปีที่แล้ว +1

      @@awindwaker4130 thanks. I realised it later. 😊👍

  • @mosquitobight
    @mosquitobight 7 หลายเดือนก่อน

    I started it wrong because I misread the square of the derivative as the second derivative. I wouldn't mind seeing the solution to that, btw.

  • @anestismoutafidis4575
    @anestismoutafidis4575 ปีที่แล้ว +1

    => Int (X^2•dx-dy)/2

  • @mcwulf25
    @mcwulf25 ปีที่แล้ว

    Is that t substitution the same as tan(theta/2) substitution?
    When I see √(1+x^2) my inclination would be to go straight in with tan(theta) but I see someone has shown how that gives us sec^3(theta) which is a nuisance to integrate.

    • @timeonly1401
      @timeonly1401 ปีที่แล้ว +1

      Let J = integral{ [sec(x)]^3} dx = integral{sec(x) [sec(x)]^2} dx .
      Let u=sec(x) & dv=[sec(x)]^2. So, du=sec(x)tan(x)dx & v=tan(x).
      J = sec(x)tan(x) - integral{sec(x) [tan(x)]^2} dx
      J = sec(x)tan(x) - integral{sec(x) ([sec(x)]^2 - 1)} dx
      J = sec(x)tan(x) - integral{[sec(x)]^3} dx + integral{sec(x)} dx
      J = sec(x)tan(x) - J + ln |sec(x) + tan(x)| + C
      Solving for J:
      J = (1/2) [ tan(x)sec(x) - J + ln | tan(x) + sec(x) | ] + C
      Then back-sub to finish.

    • @timeonly1401
      @timeonly1401 ปีที่แล้ว +1

      In the last line of equation (right after "Solving for J"), there obviously shouldn't be a J term on the RHS. :-(

    • @mcwulf25
      @mcwulf25 ปีที่แล้ว

      @@timeonly1401 Yes because it has gone to the left 👍

  • @scottleung9587
    @scottleung9587 ปีที่แล้ว +2

    I just cheated by looking at the integral table!

  • @dienosorpo
    @dienosorpo ปีที่แล้ว +1

    But my teacher always told me dy/dx is a symbol as a whole, not a damn division 😭😭😭
    I do not know anything about calculus but I feel like that's incorrect.
    The integral has a dx at the end to make it clear that you are integrating with respect to x.
    And the integral with respect to x of the derivative with respect to x is the function.
    That's why on the other side there appears a "dx" because you integrated both sides with respect of x. Not because you multiplied both sides by dx.
    Am I correct? Or not? I'm just in highschool so I could be wrong

    • @gabrijel5341
      @gabrijel5341 ปีที่แล้ว +1

      You are indeed correct, dy/dx is a symbol as a whole, but in this case you can treat it as division and it will not matter and it helps solving this equation easier.

  • @Vladimir_Pavlov
    @Vladimir_Pavlov ปีที่แล้ว +1

    I give you the third way. Integration "in parts".
    Int = ∫√(x^2+1)dx = | u=√(x^2+1) ⇒ du=x/√(x^2+1) , dv=dx ⇒v=x |
    = x*√(x^2+1)- ∫x^2dx/√(x^2+1) =x√(x^2+1)- Int +∫dx/√(x^2+1).
    | the last integral is considered tabular all over the world - "long logarithm"|
    2* Int= x√(x^2+1)+ ln(√(x^2+1) +x)+C,
    Int= (1/2) *x√(x^2+1)+ (1/2)*ln[√(x^2+1) +x]+C,
    the sign of the absolute value of the logarithm argument does not need to be set. Positivity is fulfilled .)

  • @doyourealiseiamaghost5862
    @doyourealiseiamaghost5862 ปีที่แล้ว +1

    sub x with sinh(u) for easier approach

  • @alextang4688
    @alextang4688 ปีที่แล้ว +16

    This is a standard problem in advanced level mathematics at high school. 😋😋😋😋😋😋

    • @rudra234-s1j
      @rudra234-s1j ปีที่แล้ว +1

      This is not very fast question

    • @AbouTaim-Lille
      @AbouTaim-Lille ปีที่แล้ว +1

      This is an elementary problem. First thing it is of rank 1 unlike the Bessel differential equation for example . And 2nd thing the differential y' is solvable as a function of x . And after all we are still talking about ordinary differential equations !! ..... So it can be integrated directly .

  • @roldansosabarajas1213
    @roldansosabarajas1213 ปีที่แล้ว

    Podría hablar un poco más lento por favor

  • @arnabbose5398
    @arnabbose5398 ปีที่แล้ว

    t or not to t

  • @broytingaravsol
    @broytingaravsol ปีที่แล้ว

    y=c±[x√(1+x²)+ln|x+√(1+x²)|]/2

  • @barakathaider6333
    @barakathaider6333 ปีที่แล้ว +1

    👍