The legendary A5 Putnam integral: Richard Feynman vs Weierstrass

Поділитися
Вставка
  • Опубліковано 1 гру 2024

КОМЕНТАРІ • 26

  • @violintegral
    @violintegral Рік тому +32

    Substituting x = tan(u) and then taking u --> pi/4 - u is equivalent to the substitution t = (1 - x)/(1 + x) (which implies x = (1 - t)/(1 + t)) since tan(pi/4 - u) = (1 - tan(u))/(1 + tan(u)) = (1 - x)/(1 + x) = t. Knowing this, using the substitution x = (1 - t)/(1 + t) and exploiting the beautiful symmetry within the integral is undoubtedly the most elegant solution in my mind, since it requires one less substitution and we don't have to venture into the trig world. Feynman's trick is always cool tho. I've never heard anyone call the substitution x = (1 - t)/(1 + t) the Weierstrass substitution, I thought that was just t = tan(x/2). That substitution is actually awesome tho, I used it to solve another integral on your channel by symmetry, the integral from 0 to infinity of (x - 1)/(sqrt(2^x - 1)*ln(2^x - 1)) w.r.t. x. I think a general case of this type of rational function substitution could be very useful for evaluating various definite integrals but I have yet to look into that.

  • @jyotsanabenpanchal7271
    @jyotsanabenpanchal7271 3 місяці тому

    Was looking for this, thank you.

  • @musicmakelightning
    @musicmakelightning Рік тому +3

    It's been a long time since I was in school and had to do this kind of math, but it is so cool seeing someone who has a natural instinct for it showing all the cool tricks. Makes me feel young again. Thanks for your videos.

  • @thisisntthefirsttime
    @thisisntthefirsttime Рік тому +1

    subtituting x = tan(u) and combining 1 + tan(u) to form (sin(u) + cos(u))/cos(u) allows you to apply the harmonic addition theorem for the numerator and to split the integral up by log properties.
    afterwards, you can apply log properties once again to pull out pi*ln(2)/8 and you're left with two integrals that differ by a parameter. to deal with this you can apply "king's property" and even function properties to arrive at pi*ln(2)/8.

  • @TheEternalVortex42
    @TheEternalVortex42 Рік тому +1

    Fantastic

  • @zunaidparker
    @zunaidparker Рік тому +2

    Great video. Could you do a video about the Weierstrass substitution itself? E.g., explaining the link to tan(x/2), how it was discovered, how to identify when to use it, etc.

  • @manstuckinabox3679
    @manstuckinabox3679 Рік тому +2

    No contour Integral is like the Re(the complex man)...
    Man weistrauss subs are awesome! I should look over it's "rule of thumb" in regards in how to use it.

  • @doronezri1043
    @doronezri1043 6 місяців тому

    Great video Kamal!
    How about writing 1/(1+x^2) as a geometric sum and the do integration by part (using polynomial division)? 😊

  • @sammtanX
    @sammtanX 5 місяців тому

    BRO IS STILL CALM IN HERE

  • @anthonyjohnfinch6870
    @anthonyjohnfinch6870 Місяць тому

    Hi Kamal. I’ve been following your integrals enthusiastically but I can’t access your descriptions! I can’t even see how to describe or see a list of the integrals you have solved. I’d like to see a proof of Euler’s reflection formula which doesn’t involve contour integration. This would imply integrating x^(n-1)/(x+1) without contour integration. Is this possible and, if so, how?

    • @maths_505
      @maths_505  Місяць тому

      Here's a proof for the reflection formula (2nd half of the video):
      ua-cam.com/video/5VE5kJUJFE0/v-deo.html

  • @thomasblackwell9507
    @thomasblackwell9507 Рік тому

    In my opinion, you take the win for coming up with the ways to solve it.

  • @hussainfawzer
    @hussainfawzer Рік тому

    In what device your making these videos ?
    Tablet ? Or phone ? Or computer ?

  • @holyshit922
    @holyshit922 Рік тому

    I like the second approach because it is the easiest way to calculate it in my opinion

  • @rudycummings4671
    @rudycummings4671 Рік тому +1

    Would like to see a solution using contour integration

  • @palpitate2005
    @palpitate2005 Рік тому

    Can it be solved by contour integration? (for Feyman's trick,let I(a)=int[ln(1+ax)/(1+x^2)].It is also possible.

  • @melancholic_mood_37
    @melancholic_mood_37 Рік тому

    No links in the description!

  • @lifelover9485
    @lifelover9485 3 місяці тому

    the answer should be п*ln(4)/8

  • @Ayush-yj5qv
    @Ayush-yj5qv 4 місяці тому

    Standard integral in iit