Let's Solve An Interesting Differential Equation

แชร์
ฝัง
  • เผยแพร่เมื่อ 6 ก.พ. 2025

ความคิดเห็น • 20

  • @NadiehFan
    @NadiehFan 11 หลายเดือนก่อน +3

    Another approach would be to use an integrating factor (which incidentally is a method devised by Euler). The equation to solve is
    y' + y = x²eˣ
    Multiplying both sides by eˣ we have
    y'·eˣ + y·eˣ = x²·e²ˣ
    The left hand side is now the derivative of y·eˣ so integrating both sides we get
    y·eˣ = (½x² − ½x + ¼)·e²ˣ + c₁
    and multiplying both sides by e⁻ˣ we have
    y = (½x² − ½x + ¼)·eˣ + c₁·e⁻ˣ
    and the differential equation is solved. Note that the general solution of the first order linear inhomogeneous differential equation only contains a _single_ constant. Your addition of a second constant at the end of your solution is wrong: clearly there can exist no two solutions of this differential equation which only differ by a constant.
    Alternatively, you could solve the homogeneous equation by means of its characteristic equation as you did in the video (again, a method devised by Euler) and then find a particular solution of the inhomogeneous equation by making an _Ansatz_
    yₚ = (ax² + bx + c)·eˣ
    and substituting this in the inhomogeneous equation together with its derivative
    yₚ' = (ax² + (2a + b)x + (b + c))·eˣ
    Equating coefficients we then have
    2a = 1
    2a + 2b = 0
    b + 2c = 0
    which gives a = ½, b = −½, c = ¼, so we have a particular solution
    yₚ = (½x² − ½x + ¼)·eˣ
    and together with the general solution of the homogeneous equation
    yₕ = c₁·e⁻ˣ
    this again gives the general solution y = yₕ + yₚ of the inhomogeneous equation as
    y = c₁·e⁻ˣ + (½x² − ½x + ¼)·eˣ

    • @SyberMath
      @SyberMath  11 หลายเดือนก่อน

      very good!

  • @theelk801
    @theelk801 11 หลายเดือนก่อน +4

    you can multiply both sides by e^x and then the left side becomes the derivative of y*e^x which makes it easy to solve

    • @SyberMath
      @SyberMath  11 หลายเดือนก่อน

      Great! You use an IF

  • @dariosilva85
    @dariosilva85 11 หลายเดือนก่อน +7

    You have one extra constant C on the last line which screws up the solution.

    • @lettucehelper
      @lettucehelper 11 หลายเดือนก่อน +3

      Actually, the problem is that the constant should have been added at the end of v(x) which is then multiplied by e^(-x) in the particular solution. This would, then, make the final term of the general solution ce^(-x). And this will satisfy the original equation.

    • @lettucehelper
      @lettucehelper 11 หลายเดือนก่อน +1

      Actually, I just realized that final term (ce^(-x)) is extraneous, since it can just be added to the first term.

  • @goldfing5898
    @goldfing5898 11 หลายเดือนก่อน +1

    The product of a polynomial and e^x is a again a product of a polynomial and e^x, with the polynomial having the same degree.
    So I made the Ansatz
    y(x) = (ax^2 + bx + c) * e^x
    y'(x)
    = (ax^2 + bx + c) * e^x + (2ax + b) * e^x
    = (ax^2 + (2a + b)x + (b + c)) * e^x
    The sum is
    y'(x) + y(x)
    = (2ax^2 + (2a + 2b)x + (b + 2c)) * e^x
    And shall be equal to
    x^2 * e^x = (1*x^2 + 0*x + 0) * e^x
    Comparing the coefficients yields
    2a = 1
    a = 1/2
    2a + 2b = 0
    a + b = 0
    1/2 + b = 0
    b = -1/2
    b + 2c = 0
    -1/2 + 2c = 0
    2c = 1/2
    c = 1/4
    Altogether, I get the solution function
    y(x)
    = (1/2*x^2 - 1/2*x + 1/4) * e^x
    = 1/4 * (2x^2 - 2x + 1) * e^x
    and
    y'(x)
    = 1/4 * [ (2x^2 - 2x + 1) * e^x + (4x - 2) * e^x ]
    = 1/4 * (2x^2 + 2x - 1) * e^x
    Their sum is indeed
    1/4 * 4x^2 * e^x
    = x^2 * e^x

  • @pierluigifogliamanzillo1842
    @pierluigifogliamanzillo1842 11 หลายเดือนก่อน +3

    The +c constant in the final expression of the general solution must not be there.
    Indeed if you consider the linear operator L(y) = y’+y, L(c1 e^-x) = 0, L(1/2(x2-x+1/2)e^x) = x2e^x and L(c) = c. Therefore, given the linearity of L, and assuming the reported form of y_g(x), L(y_g(x)) = x2e^x +c, which means y_g would not satisfy the given equation.

    • @lettucehelper
      @lettucehelper 11 หลายเดือนก่อน +1

      Yes, the problem is he didn't add the constant term when he solved for v(x). Of course, that makes the final term of the general equation ce^(-x), which is extraneous, since it can just be added to the first term of the general solution to form a new arbitrary constant.

  • @holyshit922
    @holyshit922 11 หลายเดือนก่อน

    First order linear equation
    We can use :
    undetermined coeffciens
    (Homogegeous part has constant coefficients and non-homogeneous part is in the form P(x)e^{λx} , where λ \in C and P(x) is polynomial,
    because of superposition we can also have linear combination of functions in mentioned form)
    We can use integrating factor or variation of parameter
    For variation of parameter we need to solve homogeneous equation first

  • @moeberry8226
    @moeberry8226 11 หลายเดือนก่อน +1

    You could have used the y particular method in the first example you just have to make sure that it is linearly independent from y homogeneous. And in this case it is. Even if it’s not you just need to multiply by x until it becomes linearly independent.

  • @scottleung9587
    @scottleung9587 11 หลายเดือนก่อน

    Nice!

  • @kingbeauregard
    @kingbeauregard 11 หลายเดือนก่อน +1

    You sure that the "+ C" belongs? We have a "+ C" on normal integrals because it's the homogeneous solution to equations of the form "y' = f(x)". But this equation isn't in that form, and the homogeneous solution is something else.

  • @Hobbitangle
    @Hobbitangle 10 หลายเดือนก่อน

    The simple substitution
    y(x)=z(x)•exp(x)
    converts the original DE into
    z'+2•z=x²
    What could be easily solved
    The particular solution can be found in form
    z(x) = x²+a•x+b
    z'(x) = 2x+a
    what implies
    x²+a•x+b+4x+2a=x²
    a=-4
    b=8
    z(x)=x²-4x+8
    The general solution is
    z(x)= C•exp(-2x)+x²-4x+8
    y(x)= C•exp(-x)+(x²-4x+8)•exp(x)
    Where the constant C must be found from the initial condition:
    y(0)=C+8

  • @giuseppemalaguti435
    @giuseppemalaguti435 11 หลายเดือนก่อน +1

    λ=-1..yo(x)=ce^(-x)...siccome -1 è diverso dall esponente di e^x,risulta semplicemente yp(x)=(ax^2+bx+c)e^x=(1/2x^2-1/2x+1/4)e^x

  • @FisicTrapella
    @FisicTrapella 11 หลายเดือนก่อน

    👌

  • @Sparky1_1
    @Sparky1_1 11 หลายเดือนก่อน

    Linear DE

  • @hetro9969
    @hetro9969 11 หลายเดือนก่อน +1

    I am from nepal