I am working with the following inhomogeneous differential equation, $$x''+x=3\cos (\omega t)$$

The general solution for this is $x(t)=x_h(t)+x_p(t)$

First step is to find $x_h(t):$

So the characteristic equation is, $$\lambda^2+0 \lambda+1=0$$ and its roots are $$\lambda =\frac{\sqrt{-4}}{2}=\frac{i\sqrt{4}}{2}=\pm i$$ So $$x_h(t)=c_1 \cos(t)+c_2 \sin(t)$$

Second step is to find $x_p(t):$

My guess will be, $$x_p(t)=A \cos(\omega t)+B \sin(\omega t)$$

Now I take the derivative of my guess. $$x_p'(t)=-A \sin(\omega t) \cdot \omega+B \cos(\omega t) \cdot \omega$$ $$x_p''(t)= -A \cos(\omega t) \cdot \omega^2-B \sin(\omega t) \cdot \omega^2$$

The I have replace it in the equation $$-A \cos(\omega t) \cdot \omega^2-B \sin(\omega t) \cdot \omega^2 + A \cos(\omega t)+B \sin(\omega t) = 3 \cos(\omega t)$$

Since we have no $\sin(\omega t)$ on the RHS the B must be $0$ on the LHS. Then $$-A \cos(\omega t) \cdot \omega^2 - A \cos(\omega t) = 3 \cos(\omega t)$$

And isolate A $$A\big(- \cos(\omega t) \cdot \omega^2-\cos(\omega t)\big)=3 \cos(\omega t)$$ $$A=\frac{3 \cos(\omega t)}{- \cos(\omega t) \cdot \omega^2-\cos(\omega t)}$$ $$A=\frac{3 \cos(\omega t)}{\cos(\omega t)\big(-\omega^2-1 \big)}$$ $$A=\frac{3}{-\omega^2-1}$$

Then replace this into our guess $$x_p(t)=\frac{3}{-\omega^2-1} \cos(\omega t)$$ $$x_p(t)=\frac{3 \cos(\omega t)}{-\omega^2-1}$$

Last the general solution is $$x(t)=c_1 \cos(t)+c_2 \sin(t)-\frac{3 \cos(\omega t)}{\omega^2-1}$$

  • $\begingroup$ Is your $x_p$ satisfies the differential equation? $\endgroup$ – user99914 Oct 21 '15 at 11:03
  • $\begingroup$ To know that I should take the derivative twice and plug it into the equation I think... $\endgroup$ – Alim Teacher Oct 21 '15 at 11:06
  • $\begingroup$ Yes, so is $x_p'' + x_p = 3\cos (\omega t)$? $\endgroup$ – user99914 Oct 21 '15 at 11:06
  • $\begingroup$ No it doesn't satisfies $\endgroup$ – Alim Teacher Oct 21 '15 at 11:12
  • $\begingroup$ Do you know in general how to guess a particular solution? There is something called variation of constant. $\endgroup$ – user99914 Oct 21 '15 at 11:19

The homogeneous solution is $$ x_h(t) = c_1\cos t + c_2\sin t $$

There are two cases:

If $\omega \ne 1$, the particular solution is $$ x_p(t) = A\cos(\omega t) + B\sin(\omega t) $$

where $A$ and $B$ are constants (not $3$, you don't know what they are yet).

If $\omega = 1$, our previous guess becomes $x_p(t) = A\cos t + B\sin t$, which is the same as the homogeneous solution. When you plug in this guess, you will get $0$ on the RHS. In order to fix it, we have to add a factor of $t$ $$ x_p(t) = At\cos t + Bt\sin t$$

  • $\begingroup$ So what do you think, should I choose $\omega \ne 1$ or $\omega =1$? $\endgroup$ – Alim Teacher Oct 21 '15 at 20:31
  • $\begingroup$ That depends on the problem. If it doesn't tell you what $\omega$ is, you have to do both $\endgroup$ – Dylan Oct 22 '15 at 0:10
  • $\begingroup$ I edited my calculations. please take a look. $\endgroup$ – Alim Teacher Oct 22 '15 at 10:05
  • $\begingroup$ Wait I got it I just solved it. $\endgroup$ – Alim Teacher Oct 22 '15 at 10:11
  • $\begingroup$ You're not done. What you have is the solution for $\omega \ne 1$. When $\omega = 1$, that solution is undefined. You need a different solution for that specific case (using the other guess). $\endgroup$ – Dylan Oct 22 '15 at 20:33

In general a second order equation of the type \begin{equation*} \partial _{t}^{2}x(t)+a\partial _{t}x(t)+bx(t)=f(t) \end{equation*} can be handled by rewriting it as a coupled set of first order equations by setting \begin{equation*} x_{1}=x,\;x_{2}=\partial _{t}x \end{equation*} Then, denoting \begin{equation*} \mathbf{x}(t)=\left( \begin{array}{c} x_{1}(t) \\ x_{2}(t) \end{array} \right) \end{equation*} one obtains the equation \begin{equation*} \partial _{t}\mathbf{x}(t)=\mathsf{A}\cdot \mathbf{x}(t)+\mathbf{y}(t) \end{equation*} where $\mathsf{A}$ is a $t$-independent $2\times 2$ matrix. In your case \begin{equation*} \partial _{t}^{2}x(t)+x(t)=3\cos (\omega t) \end{equation*} we have \begin{equation*} \mathsf{A}=\left( \begin{array}{cc} 0 & 1 \\ -1 & 0 \end{array} \right) ,\;\mathbf{y}(t)=\left( \begin{array}{c} 0 \\ 3\cos (\omega t) \end{array} \right) \end{equation*} and the solution is \begin{equation*} \mathbf{x}(t)=\exp [\mathsf{A}t]\cdot \mathbf{x}(0)+\int_{0}^{t}ds\exp [% \mathsf{A}(t-s)]\cdot \mathbf{y}(s) \end{equation*} Since \begin{equation*} \mathsf{A}^{2}=-\left( \begin{array}{cc} 1 & 0 \\ 0 & 1 \end{array} \right) =-\mathsf{I} \end{equation*} you have \begin{equation*} \exp [\mathsf{A}t]=\mathsf{I}\cos t+\mathsf{A}\sin t \end{equation*} Given \begin{equation*} \mathbf{y}(s)=\left( \begin{array}{c} 0 \\ 3\cos (\omega s) \end{array} \right) \end{equation*} you can do the $s$-integral and obtain the required result.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.