A Beautiful Cubic Diophantine Equation

Поділитися
Вставка
  • Опубліковано 19 лис 2024

КОМЕНТАРІ • 24

  • @pwmiles56
    @pwmiles56 4 місяці тому +8

    Nice problem! You can avoid the quadratic by making two substitutions
    u = a - b
    v = a+ b
    We have
    (u^2 + v^2)/2 = u^3
    Rearranging
    v^2 = u^2(2u -1)
    Therefore u divides v. Put v = ku
    k^2 u^2 = u^2(2u - 1)
    k^2 = 2u -1
    u = (k^2 + 1)/2
    k must be odd. Put k = 2c + 1
    u = 2c^2 + 2c + 1
    v = (2c + 1)(2c^2 +2c + 1)
    a = (u + v)/2 = (c + 1)(2c^2 +2c + 1)
    b = (u - v)/2 = c (2c^2 +2c + 1)

    • @MrGeorge1896
      @MrGeorge1896 4 місяці тому +1

      Yes, did the same substitutions. At the end we all get the same result but a+b und a-b felt a bit more natural to me...

  • @BurlHall
    @BurlHall 4 місяці тому +9

    Isn’t a=0, b=0 a solution as well?

  • @Chrisoikmath_
    @Chrisoikmath_ 4 місяці тому +2

    Nice to hear your voice again! I missed it! 🤩
    Awsome!

  • @bsmith6276
    @bsmith6276 4 місяці тому +2

    I like the initial substitution a-b=n, but because of the symmetry in the problem I found it easier to continue adding in its brother a+b=m.
    Then a^2+b^2=(m^2+n^2)/2 and (a-b)^3=n^3. Therefore (m^2+n^2)/2 = n^3. Solve for m^2 to get m^2=n^2*(2n-1). SInce we are working over Z we can conclude 2n-1 is an odd square itself.
    Let 2n-1=(2c+1)^2. Then n=2c^2+2c+1 and m=+/-(2c+1)*(2c^2+2c+1).
    For the positive root of m; we have a=(m+n)/2=(c+1)*(2c^2+2c+1) and b=(m-n)/2=c*(2c^2+2c+1)
    For the negative root of m; we have a=(m+n)/2=-c*(2c^2+2c+1) and b=(m-n)/2=(-c+1)*(2c^2+2c+1)
    I do think this method is far more elegant as we get the parameterizations without needing to bring in the quadratic formula.

  • @michaelschmitt2427
    @michaelschmitt2427 4 місяці тому

    This is an especially nice chapter/video. Thanks!

  • @allanmarder456
    @allanmarder456 4 місяці тому

    Let a=k*b then (k*b)^2 + b^2 = (k*b -b)^3 assume b not 0 then b= (1+k^2)/(k-1)^3 We must choose k such that b is an integer. Let k=(n+1)/n n an integer.
    Then b= 1+[(n+1)/n]^2 / 1/(n^3) simplifying b= n*3 + n*(n+1)^2 or b=n*[n^2 +(n+1)^2] a=k*b a=[(n+1)/n]*n*[n^2 +(n+1)^2] = (n+1)*[n^2 +(n+1)^2]. This was one of
    the formulas derived in the video. Similarly by letting b=k*a you can get the the other formula. I know this method is incomplete because I haven't shown that the
    substitution k=(n+1)/n is the only value for k that will produce an integer answer for a or b. I don't know how to show that. If someone has a proof please share.
    But this method does produce a valid parameter method for getting the answers.in the video even though you can't conclude they are all the answers. Just thought
    this might be interesting.

  • @scottleung9587
    @scottleung9587 4 місяці тому

    Cool - nice to hear your voice again!

  • @lathasurabhi5153
    @lathasurabhi5153 4 місяці тому +1

    How do you get these type of ideas to solve these sums Cyber

  • @rakenzarnsworld2
    @rakenzarnsworld2 4 місяці тому +1

    a = 0, b = 0

  • @yoav613
    @yoav613 4 місяці тому

    Syber is back!🎉😊

  • @SidneiMV
    @SidneiMV 4 місяці тому

    (a - b)² + 2ab = (a - b)³
    [ I ]
    a = b =>. a - b = 0
    2ab = 0 => *a = b = 0*
    [ II ]
    a ≠ b
    ..... I don't know what to do .....

  • @phill3986
    @phill3986 4 місяці тому

    😊🎉😊👍😎

  • @vladimirkaplun5774
    @vladimirkaplun5774 4 місяці тому

    Not bad at all

  • @tetsuyaikeda4319
    @tetsuyaikeda4319 4 місяці тому

    nice voice came back
    had you been cold ?

    • @SyberMath
      @SyberMath  4 місяці тому

      Change the l to vi 🤪

    • @tetsuyaikeda4319
      @tetsuyaikeda4319 4 місяці тому

      @@SyberMath really??
      terrible ! and you are safe now.good!

  • @ridaoueld5939
    @ridaoueld5939 4 місяці тому

    ?