Finding Particular Intergrals for ODE’s, Is it okay if particular integral varies

calculusordinary differential equationsproof-verification

The equation to solve : $y''-6y'+9y=e^{3x}$

This is the general solution from wolframalpha :

$y(x) = c_2 e^{3 x} x + c_1 e^{3 x} + 1/2 e^{3x} x^{2}$

This is the solution I calculated from the method of undetermined coefficients :

$y(x) = c_2 e^{3 x} x + c_1 e^{3 x} + Ax^2 e^{3x}$

This is the solution I calculated using the 'D' operator and shortcut method :

$y(x) = c_2 e^{3 x} x + c_1 e^{3 x} + e^{3x}/18$

I think I'm getting different particular integrals while using different methods because this differential equation has a family of solutions ?

But the fact that the last solution's P.I doesn't even have an 'x' term concerns me.

All of them are correct, am I right ?

Best Answer

Testing the solution $y(x) = b e^{3 x} x + a e^{3 x} + c x^2 e^{3x}$, we get

$$2c e^{3 x} \ne e^{3x}$$

This is not a solution.


Testing the solution $y(x) = b e^{3 x} x + a e^{3 x} + e^{3x}/18$, we get

$$0 \ne e^{3x}$$

This is not a solution.


Testing the solution, $y(x) = c_2 e^{3 x} x + c_1 e^{3 x} + 1/2 e^{3x} x^{2}$, we get

$$e^{3x} = e^{3x}$$

This is a solution.


Now, lets use Undetermined Coefficients to solve $y''-6y'+9y=e^{3x}$.

For the homogeneous solution, we have $ m^2 - 6 m + 9 = 0 \implies m_{1,2} = 3$, hence

$$y_h(x) = e^{3x}( c_1 + c_2 x)$$

Because our homogeneous solution, $e^{3x}(c_1 + c_2 x)$ shares the RHS $e^{3x}$, we choose $$y_p(x) = a x^2 e^{3x}$$

Substituting, we have

$$y_p(x)''-6y_p(x)'+9y_p(x) = 2 a e^{3x} =e^{3x} \implies a = \dfrac{1}{2}$$

We can now write

$$y(x) = y_h(x) + y_p(x) = e^{3x}( c_1 + c_2 x + \dfrac{1}{2} x^2)$$

Related Question