Solve y''+y=cot(x). I know I asked a similar, I had the signs incorrect, my bad. Ok from here I have the characteristic equation is r^2+1=0 which gives r=i,-i. this give y1=cosx and y2=sinx.....right? Where do I go from here?
Like last time, I'm pretty sure you'll need to use variation of parameters. I haven't come across a solution to an equation like this that did not involve VoP, other than (maybe) power series.
Apparently we DID learn it but he never ever gave it a name in class. I had to find it in the book and it's basically like what he did in examples. It was right after undetermined coefficients. I'm sorry I didn't realize that yesterday!
That's okay! First thing you do is check the fundamental solutions \(\sin x\) and \(\cos x\) are linearly independent. Do you know how to do that with the Wronskian?
[cosx sinx -sinx cosx] and then take the determinant?
and the determinant=1 which isn't 0 so they are linearly independent....?
Right.
Now you use the formulas for the new-parameter solution, \[u_1=-\int\frac{y_2g(t)}{W(y_1,y_2)}~dt~~~~\text{and}~~~~u_2=\int\frac{y_2g(t)}{W(y_1,y_2)}~dt\] where \(y_1=\cos x\), \(y_2=\sin x\), \(W(y_1,y_2)\) is the Wronskian, and \(g(t)=\cot x\). The final solution will be \[y=u_1y_1+u_2y_2\]
oh man....the answer's gonna be ugly isn't it?
Surprisingly, no! Well, maybe a little, but one of those integrals will simplify to something really nice.
should that be y1 in the formula for u1?
Nope, \(u_1\) is right, \(u_2\) is wrong: \[u_1=-\int\frac{y_2g(t)}{W(y_1,y_2)}~dt~~~~\text{and}~~~~u_2=\int\frac{\color{red}{y_1}g(t)}{W(y_1,y_2)}~dt\] Thanks for catching that.
so.....sorry I'm just making sure I have all steps correct.....u1 should become -sin(x)...and u2= cos(x) +log(sin(x/2))-log(cos(x/2))
and the final, general solution will be c1cosx +c2sinx + (u1y1 +u2y2)?
\(u_1\) is correct. I don't have any paper at hand so I'll check your work with the second integral: \[\begin{align*}u_2&=\int\cos x\cot x~dx\\&=\cot x\sin x+\int\csc^2x\sin x~dx&\text{integrating by parts}\\ &=\cos x+\int\csc x~dx\\ &=\cos x-\ln\left|\csc x+\cot x\right| \end{align*}\] With some identities, you'd get \[u_2=\cos x+\ln\left|\sin\frac{x}{2}\right|-\ln\left|\cos\frac{x}{2}\right|\] So yes, you're right. And yes, the general solution would be as you wrote it.
AWESOME!!!! thank you so so so much!
You're welcome!
Join our real-time social learning platform and learn together with your friends!