A Seriesly Cool Problem!

Поделиться
HTML-код
  • Опубликовано: 4 фев 2025

Комментарии • 55

  • @RandomBurfness
    @RandomBurfness 3 месяца назад +63

    I suppose it is left as an exercise to the viewer to prove that each partial sum is indeed equal to [(k+2)!-1]/[(k+2)!]?

    • @ianmathwiz7
      @ianmathwiz7 3 месяца назад +19

      It's pretty easy to do with induction.

    • @taterpun6211
      @taterpun6211 3 месяца назад +12

      1/k!(k+2) = (k+1)/(k+2)! = (k+2-1)/(k+2)!
      = 1/(k+1)! - 1/(k+2)!
      The sum terminates, leaving S_n = 1 - 1/(n+2)!

    • @yoyoezzijr
      @yoyoezzijr 3 месяца назад +1

      You can instantly see it can be proved by induction by just verifying that it still holds for k+1.

    • @henry-db
      @henry-db 3 месяца назад +1

      I was really disappointed when he didn't end up proving that equallity, mainly because, as @iammathwiz7 said, it's pretty easy to do with induction.

    • @lukevideckis2260
      @lukevideckis2260 3 месяца назад +1

      ah the good old telescoping sum

  • @tavantrai3968
    @tavantrai3968 3 месяца назад +39

    what about we multiply n+1 to above and below, we can make it into 1/(n+1)! - 1/(n+2)!, then the equation equal to (e-1)-(e-1-1)=1

    • @blueslime5855
      @blueslime5855 3 месяца назад +4

      This is what I did!

    • @bsmith6276
      @bsmith6276 3 месяца назад +5

      Even better is that doing that we done even need to know the classic convergence to e; now it is just a very simple telescoping series where everything cancels out except the first 1, so the sum is 1.

    • @blueslime5855
      @blueslime5855 3 месяца назад +1

      @@bsmith6276 ohh I see, so all terms for n≥2 cancel and we just have 1/1!
      cool, even cleaner.
      I'm surprised he didn't see these ways to solve it, but I still think his is interesting as well

    • @krabkrabkrab
      @krabkrabkrab 3 месяца назад +1

      I did that too. Quicker, and without any blackboard.

  • @TheSabian321
    @TheSabian321 3 месяца назад +10

    Of course, the answer is the mathematical constant Wau. It's truly everywhere.

  • @pragyanpranay3681
    @pragyanpranay3681 3 месяца назад +6

    Bro did not induct 😭

  • @ianmathwiz7
    @ianmathwiz7 3 месяца назад +10

    8:05 This formula is also very easy to prove via induction.

  • @steve112285
    @steve112285 3 месяца назад +11

    An alternative approach: Since the series for x*e^x centered at 0 converges on (-infinity,infinity), and [0,1] is contained in this interval, we can interchange the order of integration and summation for the power series. Applying power rule for antiderivatives at the end, and evaluating from 0 to 1, we have
    Integrate[x*e^x,{x,0,1}]
    =Integrate[x*Sum[x^n/n!,{n,0,infinity}],{x,0,1}]
    =Integrate[Sum[x^(n+1)/n!,{n,0,infinity}],{x,0,1}]
    =Sum[Integrate[x^(n+1)/n!,{x,0,1}],{n,0,infinity}]
    =Sum[1/(n!*(n+2)),{n,0,infinity}]
    Using integration by parts, the value of the original integral is 1, so the series has the same value.

  • @ars_analytica
    @ars_analytica 3 месяца назад +3

    This collapses to a telescoping series after multiplying by n+1 on the top and bottom and then writing n+1 as n+2-1. Differentiating under the sum can also work though it is more involved. There are at least 4 ways of doing this. Either cancel the n+2 in the denominator by differentiating the sum with a t^(n+2) in the numerator and use t=0 as a boundary condition. Similarly we can multiply by n+1 on the top and bottom and then introduce it by differentiating t^(n+1)/(n+2)! as the summand. An exponential can also be employed though it requires more work for no added benefit except perhaps to practise the technique.

  • @aahannayak7746
    @aahannayak7746 3 месяца назад +6

    Here's another method, using the expansion of e^x. e^x=summation x^n/n! till n tends to infinity. Multiply x on both sides and integrate. Put limits from 0 to 1 and you have the answer.

  • @mamallama7375
    @mamallama7375 3 месяца назад +3

    Could you use the maclaurin series of xe^x and integrate?

  • @chaosredefined3834
    @chaosredefined3834 3 месяца назад +2

    1/n!(n+2) = (n+1)/(n+2)! = (n+2 - 1)/(n+2)! = (n+2)/(n+2)! - 1/(n+2)! = 1/(n+1)! - 1/(n+2)!
    So, the partial series from n=0 to n=k can be rewritten as (1/1! - 1/2!) + (1/2! - 1/3!) + (1/3! - 1/4!) + ... + (1/(k+1)! - 1/(k+2)!). And this is a telescoping series, so we can cancel out a lot of common terms, giving us 1/1! - 1/(k+2)!.

  • @damianmirza8332
    @damianmirza8332 3 месяца назад +4

    10:10 ... I know that 1 is not prime...

  • @MrMctastics
    @MrMctastics 3 месяца назад +4

    Index magic:
    = n+1 /(n+2)! from 0
    = n/(n+2)! from 0+ 1/(n+2)! from 0
    = n/(n+2)! from 0+ (e-1)
    which is equal to from 1 so shift
    =n-1/(n+1)! from 1 + (e-1)
    =n/(n+1)! from 1 - (e-1) + (e-1)
    =n/(n+1)! from 0
    which is equal to from 1 so shift
    = (n-1)/n! from 1
    = 1/(n-1)! -1/n!
    everything cancels except for 1

  • @gametimewitharyan6665
    @gametimewitharyan6665 3 месяца назад +2

    I was literally solving this question a few minutes ago and now this popped up on my feed lmao

  • @DavidWang-q9z
    @DavidWang-q9z 3 месяца назад

    That was amazing dude. All it was though was me and Peter from Free University of Berlin in a blink of a second without being an absolute psychopath though. There was even a German Lady taking a practice test too and that didn't do anything.

  • @shadowreaper-gq6bz
    @shadowreaper-gq6bz 3 месяца назад +1

    we can transform "1/(n+2) = int (0,1) of x^(n+1)" it gives the integral from (0,1) of xe^x which gives "1"

    • @rosiefay7283
      @rosiefay7283 2 месяца назад

      But that's irrelevant. We have a sum here, not an integral.

    • @shadowreaper-gq6bz
      @shadowreaper-gq6bz 2 месяца назад

      @@rosiefay7283 there are some sums that had gamma function like sum from 1 to infinity of (-1)^n.gamma²(n)/gamma(2n) , and it can be solved by the beta function which is already an integral

  • @supermario2112
    @supermario2112 3 месяца назад +5

    It's also solvable through Taylor series for exp(x) and the generating functions method (the value of the series sum ends up being exactly (exp(x)/x - 1/x - 1)' evaluated at x=1)

  • @g4_684
    @g4_684 3 месяца назад +2

    I read pokemon instead of problem. Still excited

  • @gamitorre2241
    @gamitorre2241 3 месяца назад +1

    cool way to approximate 1😮

  • @alexandersanchez9138
    @alexandersanchez9138 3 месяца назад +1

    This series telescopes.

  • @mr.surlaw4239
    @mr.surlaw4239 3 месяца назад +1

    Oh fuck yeah

  • @analog_dreamer
    @analog_dreamer 3 месяца назад +3

    this is seriesly cool!

  • @VincentKok458
    @VincentKok458 2 месяца назад

    When will there be a new flammy wood vid?

  • @elnopo4242
    @elnopo4242 3 месяца назад

    In would be cool to prove that the pattern found is indeed correct for each k € N

  • @themightyripples6582
    @themightyripples6582 3 месяца назад +2

    Getting concerned... why say "good morning" like a regular person... you're better than that. You're 2 seconds of silly often hype me up for the math and I'm getting concerned and sad

  • @joyboricua3721
    @joyboricua3721 3 месяца назад

    🔥

  • @varionmori709
    @varionmori709 2 месяца назад

    Did you mean that one has been a prime number for ages, or one has been a PRIME number for ages? I'm just making sure I write it down in my notes the correct way. 😎

  • @mastershooter64
    @mastershooter64 3 месяца назад +3

    If it's in physics, it's convergent!

  • @LinkenCV
    @LinkenCV 3 месяца назад +1

    (n+1)/(n+1), -> (n+1)/(n+2)!, ->(n+1+1-1)/(n+2)! -> 1/(n+1)! - 1/(n+2)! ->reindexing -> 1+sum-sum ->=1

  • @projectpiano5231
    @projectpiano5231 3 месяца назад +3

    If it's +1 instead of +2, the result is e-1 and if it's +3 instead of +2, the result is e-2. I haven't looked into it much but I'm curious if anybody knows why off the top of their head

    • @steve112285
      @steve112285 3 месяца назад +2

      Here's one way to understand that, copying from another comment I made: Since the series for x*e^x centered at 0 converges on (-infinity,infinity), and [0,1] is contained in this interval, we can interchange the order of integration and summation for the power series. Applying power rule for antiderivatives at the end, and evaluating from 0 to 1, we have
      Integrate[x*e^x,{x,0,1}]
      =Integrate[x*Sum[x^n/n!,{n,0,infinity}],{x,0,1}]
      =Integrate[Sum[x^(n+1)/n!,{n,0,infinity}],{x,0,1}]
      =Sum[Integrate[x^(n+1)/n!,{x,0,1}],{n,0,infinity}]
      =Sum[1/(n!*(n+2)),{n,0,infinity}]
      Using integration by parts, the value of the original integral is 1, so the series has the same value.
      If you wanted n+3 instead of n+2, integrate x^2*e^x instead. etc.

  • @rosiefay7283
    @rosiefay7283 2 месяца назад

    Yes, you did some numerical work to evaluate some partial sums, and you spotted a pattern, but you should have shown us the proof. All this is no good without the proof.

  • @fengshengqin6993
    @fengshengqin6993 3 месяца назад +2

    I dont understand ,why dont we just integrate x*e^x from 0 to 1 , and get result 1 ?

  • @inyobill
    @inyobill 2 месяца назад

    That's pretty

  • @SuperSilver316
    @SuperSilver316 3 месяца назад +1

    The series is just x*e^x that we are integrating from 0 to 1

  • @aweebthatlovesmath4220
    @aweebthatlovesmath4220 3 месяца назад

    Yes daddy more

  • @fengshengqin6993
    @fengshengqin6993 3 месяца назад +1

    e^x=Sum(x^n/n!) ,two side times x and integrate from 0 to 1 ,easy to get the result ,is 1 . Why dont just do it like this way?

  • @objectified2763rebooted
    @objectified2763rebooted 3 месяца назад +1

    first