2.4 HW 3

  2.4.1 Section 2.1, problem 11
  2.4.2 Section 2.2.1, problem 6 (page 144, complex roots)
  2.4.3 section 2.2.2, problem 6 (page 149, equal roots)
  2.4.4 Section 2.4, problem 6 (page 156, Variation of parameters)
  2.4.5 Section 2.5, problem 14 (page 164, Guessing method)
  2.4.6 Key solution for HW 3
PDF (letter size)
PDF (legal size)

2.4.1 Section 2.1, problem 11

   2.4.1.1 Part a
   2.4.1.2 Part b
   2.4.1.3 Part c
   2.4.1.4 Part d

Let \(y_{1}\left ( t\right ) =t^{2}\) and \(y_{2}\left ( t\right ) =t\left \vert t\right \vert \)

1.
Show that \(y_{1},y_{2}\) are linearly dependent (L.D.) on the interval \(0\leq t\leq 1\)
2.
Show that \(y_{1},y_{2}\) are linearly independent (L.I.) on the interval \(-1\leq t\leq 1\)
3.
Show that \(W\left [ y_{1},y_{2}\right ] \left ( t\right ) \) is identically zero.
4.
Show that \(y_{1},y_{2}\) can never be two solutions of (3) which is \(y^{\prime \prime }+p\left ( t\right ) y^{\prime }+q\left ( t\right ) y=0\), on the interval \(-1<t<1\) if both \(p,q\) are continuous in this interval.

Solution

2.4.1.1 Part a

On the interval \(0\leq t\leq 1\), then \(\left \vert t\right \vert =t\) since \(t\) is positive. Hence \(y_{2}\left ( t\right ) =t^{2}\), which is the same as \(y_{1}\left ( t\right ) =t^{2}\). Therefore they are linearly dependent (same solution). In other words, \(y_{1}\left ( t\right ) =c_{1}y_{2}\left ( t\right ) \) where \(c_{1}=1.\)

2.4.1.2 Part b

When \(t\leq 0\) now \(y_{2}\left ( t\right ) =-t^{2}\). Hence we have \(y_{1}=y_{2}\) for \(0\leq t\leq 1\) and \(y_{1}=-y_{2}\) for \(-1\leq t<0\). Therefore it is not possible to find the same constant \(c\) such that \(y_{1}=cy_{2}\) which will work for all \(t\) regions. This implies that \(y_{1}\left ( t\right ) \) and \(y_{2}\left ( t\right ) \) are linearly independent on \(-1\leq t\leq 1\).

2.4.1.3 Part c

\[ W\left [ y_{1},y_{2}\right ] \left ( t\right ) =\begin{vmatrix} y_{1} & y_{2}\\ y_{1}^{\prime } & y_{2}^{\prime }\end{vmatrix} =y_{1}y_{2}^{\prime }-y_{2}y_{1}^{\prime }\] If \(W\left ( t\right ) =0\) is in some region or at some point, then it must be zero anywhere. Therefore let us pick the interval \(0\leq t\leq 1\) to calculate \(W\left ( t\right ) \). This way we avoid having to deal with the \(\left \vert t\right \vert \) when taking derivatives since on this interval, \(y_{1}=t^{2}\) and also \(y_{2}=t^{2}\). Now \(W\left ( t\right ) \) becomes\begin{align*} W\left ( t\right ) & =t^{2}\left ( 2t\right ) -t^{2}\left ( 2t\right ) \\ & =0 \end{align*}

Therefore \(W\left ( t\right ) =0\) everywhere.

2.4.1.4 Part d

Since \(p,q\) are continuous on \(-1<t<1\), then by uniqueness theorem, we know there are two fundamental solutions \(y_{1},y_{2}\), which must be linearly independent that their linear combination give the general solution \(y\left ( t\right ) =c_{1}y_{1}\left ( t\right ) +c_{2}y_{2}\left ( t\right ) \).

But from part(b) above we found that the given functions \(y_{1},y_{2}\) are not linearly independent on \(-1<t<1\), hence these can never be the fundamental solutions to \(y^{\prime \prime }+p\left ( t\right ) y^{\prime }+q\left ( t\right ) y=0\).

2.4.2 Section 2.2.1, problem 6 (page 144, complex roots)

Solve \(y^{\prime \prime }+2y^{\prime }+5y=0\) with \(y\left ( 0\right ) =0,y^{\prime }\left ( 0\right ) =2\)

Solution

Let \(y=e^{\lambda t}\). Substituting in the above ODE gives\begin{align*} \lambda ^{2}e^{\lambda t}+2\lambda e^{\lambda t}+5e^{\lambda t} & =0\\ e^{\lambda t}\left ( \lambda ^{2}+2\lambda +5\right ) & =0 \end{align*}

Since \(e^{\lambda t}\neq 0\), the above simplifies to \(\lambda ^{2}+2\lambda +5=0\). The roots are \(\lambda =\frac{-b}{2a}\pm \frac{1}{2a}\sqrt{b^{2}-4ac}=\frac{-2}{2}\pm \frac{1}{2}\sqrt{4-4\left ( 5\right ) }\) or \(\lambda =-1\pm \frac{1}{2}\sqrt{-16}\). Hence \[ \lambda =-1\pm 2i \] Therefore the general solution is linear combination of\begin{align*} y\left ( t\right ) & =c_{1}e^{\lambda _{1}t}+c_{2}e^{\lambda _{2}t}\\ & =c_{1}e^{\left ( -1+2i\right ) t}+c_{2}e^{\left ( -1-2i\right ) t}\\ & =e^{-t}\left ( c_{1}e^{2it}+c_{2}e^{-2it}\right ) \end{align*}

But \(c_{1}e^{2it}+c_{2}e^{-2it}\) can be rewritten, using Euler relation, as \(C_{1}\cos 2t+C_{2}\sin 2t\). The above solution becomes\begin{equation} y\left ( t\right ) =e^{-t}\left ( C_{1}\cos 2t+C_{2}\sin 2t\right ) \tag{1} \end{equation} \(C_{1},C_{2}\) are now found from initial conditions. At \(t=0\)\[ 0=C_{1}\] The solution (1) simplifies to\begin{equation} y\left ( t\right ) =C_{2}e^{-t}\sin 2t \tag{2} \end{equation} Taking time derivative gives\[ y^{\prime }\left ( t\right ) =C_{2}\left ( -e^{-t}\sin 2t+2e^{-t}\cos 2t\right ) \] At \(t=0\) the above becomes\begin{align*} 2 & =2C_{2}\\ C_{2} & =1 \end{align*}

Substituting the above in (2) gives the final general solution\[ y\left ( t\right ) =e^{-t}\sin 2t \]

2.4.3 section 2.2.2, problem 6 (page 149, equal roots)

Solve the following initial-value problems  \(y^{\prime \prime }+2y^{\prime }+y=0\) with \(y\left ( 2\right ) =1,y^{\prime }\left ( 2\right ) =-1\)

Solution

Let \(y=e^{\lambda t}\). Substituting in the above ODE gives\begin{align*} \lambda ^{2}e^{\lambda t}+2\lambda e^{\lambda t}+e^{\lambda t} & =0\\ e^{\lambda t}\left ( \lambda ^{2}+2\lambda +1\right ) & =0 \end{align*}

Since \(e^{\lambda t}\neq 0\), the above simplifies to \(\lambda ^{2}+2\lambda +1=0\) or \(\left ( \lambda +1\right ) ^{2}=0\). Hence there is a double root \(\lambda =-1\). One fundamental solution is \[ y_{1}=e^{-t}\] To find the second solution, reduction of order is used. Let the second solution be \begin{align} y_{2}\left ( t\right ) & =y_{1}\left ( t\right ) u\left ( t\right ) \nonumber \\ & =e^{-t}u \tag{1} \end{align}

Hence \begin{align} y_{2}^{\prime } & =-e^{-t}u+e^{-t}u^{\prime }\tag{2}\\ y_{2}^{\prime \prime } & =e^{-t}u-e^{-t}u^{\prime }-e^{-t}u^{\prime }+e^{-t}u^{\prime \prime } \tag{3} \end{align}

Substituting (1,2,3) into the ODE gives (since \(y_{2}\) is assumed to be a solution)\begin{align*} \left ( e^{-t}u-e^{-t}u^{\prime }-e^{-t}u^{\prime }+e^{-t}u^{\prime \prime }\right ) +2\left ( -e^{-t}u+e^{-t}u^{\prime }\right ) +\left ( e^{-t}u\right ) & =0\\ \left ( u-u^{\prime }-u^{\prime }+u^{\prime \prime }\right ) +2\left ( -u+u^{\prime }\right ) +u & =0\\ u^{\prime \prime }-2u^{\prime }+u-2u+2u^{\prime }+u & =0\\ u^{\prime \prime } & =0 \end{align*}

Hence the solution is \(u=C_{1}t+C_{2}\). Therefore from (1) the second solution is\begin{align*} y_{2}\left ( t\right ) & =y_{1}\left ( t\right ) u\left ( t\right ) \\ & =e^{-t}\left ( C_{1}t+C_{2}\right ) \end{align*}

Therefore the general solution is \begin{align*} y\left ( t\right ) & =C_{3}y_{1}+C_{4}y_{2}\\ & =C_{3}e^{-t}+C_{4}e^{-t}\left ( C_{1}t+C_{2}\right ) \end{align*}

Combining constants gives\begin{align*} y\left ( t\right ) & =C_{3}e^{-t}+e^{-t}\left ( C_{1}t+C_{2}\right ) \\ & =\left ( C_{3}+C_{2}\right ) e^{-t}+C_{1}te^{-t} \end{align*}

Let \(A=\left ( C_{3}+C_{2}\right ) ,B=C_{1}\), then the final solution is\begin{equation} y\left ( t\right ) =Ae^{-t}+Bte^{-t} \tag{4} \end{equation} Now \(A,B\) are found from initial conditions \(y\left ( 2\right ) =1,y^{\prime }\left ( 2\right ) =-1\). First initial condition gives from (4)\begin{equation} 1=Ae^{-2}+2Be^{-2} \tag{5} \end{equation} Taking derivative of (4) gives\[ y^{\prime }\left ( t\right ) =-Ae^{-t}+B\left ( e^{-t}-te^{-t}\right ) \] Applying second initial condition on the above gives\begin{align} -1 & =-Ae^{-2}+B\left ( e^{-2}-2e^{-2}\right ) \nonumber \\ & =-Ae^{-2}-Be^{-2} \tag{6} \end{align}

Now we need to solve (5,6) for (\(A,B\)). Adding (5,6) gives\[ 0=Be^{-2}\] Hence \(B=0\). Therefore from (5) we can now solve for \(A\)\begin{align*} 1 & =Ae^{-2}\\ A & =e^{2} \end{align*}

Hence (4) now becomes\begin{align*} y\left ( t\right ) & =e^{2}e^{-t}\\ & =e^{2-t} \end{align*}

2.4.4 Section 2.4, problem 6 (page 156, Variation of parameters)

Solve the following initial-value problems  \(y^{\prime \prime }+4y^{\prime }+4y=t^{\frac{5}{2}}e^{-2t}\) with \(y\left ( 0\right ) =0,y^{\prime }\left ( 0\right ) =0\)

Solution

The first step is to solve the homogenous ODE \(y^{\prime \prime }+4y^{\prime }+4y=0\).  The characteristic equation is \begin{align*} \lambda ^{2}+4\lambda +4 & =0\\ \left ( \lambda +2\right ) \left ( \lambda +2\right ) & =0 \end{align*}

Hence a double root at \(\lambda =-2\). The first solution is \(y_{1}=e^{=2t}\). Therefore the second solution is \(y_{2}=te^{-2t}\) (obtained using reduction of order as was done in the above problem with equal roots). Therefore the homogenous \(y_{h}\left ( t\right ) \) is \[ y_{h}\left ( t\right ) =C_{1}e^{-2t}+C_{2}te^{-2t}\] To find the particular solution \(y_{p}\left ( t\right ) \), Variation of parameters will be used. Assuming the particular solution is\[ y_{p}\left ( t\right ) =u_{1}\left ( t\right ) y_{1}\left ( t\right ) +u_{2}\left ( t\right ) y_{2}\left ( t\right ) \] Where \begin{equation} u_{1}\left ( t\right ) =-\int \frac{y_{2}\left ( t\right ) f\left ( t\right ) }{W\left ( t\right ) }dt \tag{1} \end{equation} And\begin{equation} u_{2}\left ( t\right ) =\int \frac{y_{1}\left ( t\right ) f\left ( t\right ) }{W\left ( t\right ) }dt \tag{2} \end{equation} Where in the above \(f\left ( t\right ) =t^{\frac{5}{2}}e^{-2t}\) and \(y_{1}=e^{=2t},y_{2}=te^{-2t}\). We now need to find \(W\left ( t\right ) \)\begin{align*} W\left ( t\right ) & =\begin{vmatrix} y_{1} & y_{2}\\ y_{1}^{\prime } & y_{2}^{\prime }\end{vmatrix} \\ & =y_{1}y_{2}^{\prime }-y_{2}y_{1}^{\prime }\\ & =e^{=2t}\left ( e^{-2t}-2te^{-2t}\right ) +2te^{-2t}e^{-2t}\\ & =e^{-4t}-2te^{-4t}+2te^{-4t}\\ & =e^{-4t} \end{align*}

Therefore (1) becomes\begin{align*} u_{1}\left ( t\right ) & =-\int \frac{te^{-2t}t^{\frac{5}{2}}e^{-2t}}{e^{-4t}}dt\\ & =-\int t^{\frac{5}{2}+1}dt\\ & =-\int t^{\frac{7}{2}}dt\\ & =-\frac{t^{\frac{9}{2}}}{\frac{9}{2}}\\ & =\frac{-2}{9}t^{\frac{9}{2}} \end{align*}

And (2) becomes\begin{align*} u_{2}\left ( t\right ) & =\int \frac{e^{-2t}t^{\frac{5}{2}}e^{-2t}}{e^{-4t}}dt\\ & =\int t^{\frac{5}{2}}dt\\ & =\frac{t^{\frac{7}{2}}}{\frac{7}{2}}\\ & =\frac{2}{7}t^{\frac{7}{2}} \end{align*}

Since \(y_{p}\left ( t\right ) =u_{1}\left ( t\right ) y_{1}\left ( t\right ) +u_{2}\left ( t\right ) y_{2}\left ( t\right ) \), then using the above results we obtain the particular solution\begin{align*} y_{p}\left ( t\right ) & =\left ( \frac{-2}{9}t^{\frac{9}{2}}\right ) e^{-2t}+\left ( \frac{2}{7}t^{\frac{7}{2}}\right ) te^{-2t}\\ & =e^{-2t}\left ( \frac{-2}{9}t^{\frac{9}{2}}+\frac{2}{7}t^{\frac{9}{2}}\right ) \\ & =\frac{4}{63}e^{-2t}t^{\frac{9}{2}} \end{align*}

Since \(y\left ( t\right ) =y_{h}\left ( t\right ) +y_{p}\left ( t\right ) \) then the final solution is\begin{align} y\left ( t\right ) & =\left ( C_{1}e^{-2t}+C_{2}te^{-2t}\right ) +\frac{4}{63}e^{-2t}t^{\frac{9}{2}}\nonumber \\ & =e^{-2t}\left ( C_{1}+C_{2}t+\frac{4}{63}t^{\frac{9}{2}}\right ) \tag{3} \end{align}

Now initial conditions are applied to find \(C_{1},C_{2}\). From \(y\left ( 0\right ) =0\), then (3) becomes\[ 0=C_{1}\] Hence the solution (3) simplifies to\begin{equation} y\left ( t\right ) =e^{-2t}\left ( C_{2}t+\frac{4}{63}t^{\frac{9}{2}}\right ) \tag{4} \end{equation} Taking derivatives\begin{align*} y^{\prime }\left ( t\right ) & =-2e^{-2t}\left ( C_{2}t+\frac{4}{63}t^{\frac{9}{2}}\right ) +e^{-2t}\left ( C_{2}+\left ( \frac{4}{63}\right ) \left ( \frac{7}{2}\right ) t^{\frac{7}{2}}\right ) \\ & =-2e^{-2t}\left ( C_{2}t+\frac{4}{63}t^{\frac{9}{2}}\right ) +e^{-2t}\left ( C_{2}+\frac{2}{9}t^{\frac{7}{2}}\right ) \end{align*}

Applying the second BC \(y^{\prime }\left ( 0\right ) =0\) to the above gives \[ 0=C_{2}\] The solution (4) now reduces to\[ y\left ( t\right ) =\frac{4}{63}t^{\frac{9}{2}}e^{-2t}\] Which is just the particular solution. This makes sense, since both initial conditions are zero, then the homogenous solution will be zero.

2.4.5 Section 2.5, problem 14 (page 164, Guessing method)

Find the particular solution for \(y^{\prime \prime }+2y^{\prime }=1+t^{2}+e^{-2t}\)

Solution

The first step is to solve the homogeneous solution \(y_{h}\left ( t\right ) \) of the ODE \(y^{\prime \prime }+2y^{\prime }=0\).  Let \(u=y^{\prime }\). Then the ODE becomes\[ u^{\prime }+2u=0 \] The integrating factor is \(I=e^{\int 2dt}=e^{2t}\). The above becomes\begin{align*} \frac{d}{dt}\left ( ue^{2t}\right ) & =0\\ ue^{2t} & =C_{1}\\ u & =C_{1}e^{-2t} \end{align*}

But \(y^{\prime }=u\). Integrating gives \begin{align*} y_{h}\left ( t\right ) & =\int C_{1}e^{-2t}dt+C_{2}\\ & =\frac{-1}{2}C_{1}e^{-2t}+C_{2}\\ & =C_{3}e^{-2t}+C_{2} \end{align*}

Hence the fundamental solutions are \begin{align*} y_{1} & =e^{-2t}\\ y_{2} & =1 \end{align*}

We now go back to the original ODE and find the particular solution \(y_{p}\). Since the RHS is \(p\left ( t\right ) +e^{-2t}\) where \(p\left ( t\right ) =1+t^{2}\), we can use linearity and find particular solution \(y_{p_{1}}\left ( t\right ) \) associated with \(p\left ( t\right ) \) only and then find \(y_{p_{2}}\left ( t\right ) \) associated with \(e^{-2t}\) only and then add them together to obtain \(y_{p}\left ( t\right ) \). In other words\[ y_{p}\left ( t\right ) =y_{p_{1}}\left ( t\right ) +y_{p_{2}}\left ( t\right ) \] To find \(y_{p_{1}}\left ( t\right ) \) associated with \(1+t^{2}\) we guess \(y_{p_{1}}\left ( t\right ) =C_{0}+C_{1}t+C_{2}t^{2}\). But because the ODE is missing the \(y\) term in it, then we have to multiply this guess by an extra \(t\). Therefore it becomes \[ y_{p_{1}}\left ( t\right ) =t\left ( C_{0}+C_{1}t+C_{2}t^{2}\right ) \] To find \(y_{p_{2}}\left ( t\right ) \) associated with \(e^{-2t}\) we guess \(y_{p_{2}}=Ae^{-2t}\). But because \(e^{-2t}\) is also a fundamental solution of the homogenous solution found above, we have to again adjust this and multiply the guess by \(t\). Hence it becomes\[ y_{p_{2}}\left ( t\right ) =Ate^{-2t}\] Therefore the full guess for particular solution becomes\begin{align} y_{p}\left ( t\right ) & =y_{p_{1}}\left ( t\right ) +y_{p_{2}}\left ( t\right ) \nonumber \\ & =t\left ( C_{0}+C_{1}t+C_{2}t^{2}\right ) +Ate^{-2t}\nonumber \\ & =tC_{0}+C_{1}t^{2}+C_{2}t^{3}+Ate^{-2t}\tag{1A} \end{align}

Now\begin{equation} y_{p}^{\prime }\left ( t\right ) =C_{0}+2C_{1}t+3C_{2}t^{2}+Ae^{-2t}-2Ate^{-2t}\tag{1} \end{equation} And\begin{equation} y_{p}^{\prime \prime }\left ( t\right ) =2C_{1}+6C_{2}t-2Ae^{-2t}-2Ae^{-2t}+4Ate^{-2t}\tag{2} \end{equation} Substituting (1,2) into LHS of \(y^{\prime \prime }+2y^{\prime }=1+t^{2}+e^{-2t}\) gives\begin{align*} \left ( 2C_{1}+6C_{2}t-2Ae^{-2t}-2Ae^{-2t}+4Ate^{-2t}\right ) +2\left ( C_{0}+2C_{1}t+3C_{2}t^{2}+Ae^{-2t}-2Ate^{-2t}\right ) & =1+t^{2}+e^{-2t}\\ 2C_{0}+2C_{1}+4tC_{1}+6tC_{2}-2Ae^{-2t}+6t^{2}C_{2} & =1+t^{2}+e^{-2t}\\ e^{-2t}\left ( -2A\right ) +t\left ( 4C_{1}+6C_{2}\right ) +t^{2}\left ( 6C_{2}\right ) +\left ( 2C_{0}+2C_{1}\right ) & =1+t^{2}+e^{-2t} \end{align*}

Comparing coefficients gives\begin{align*} -2A & =1\\ 4C_{1}+6C_{2} & =0\\ 6C_{2} & =1\\ 2C_{0}+2C_{1} & =1 \end{align*}

Solving gives \(A=-\frac{1}{2},C_{0}=\frac{3}{4},C_{1}=-\frac{1}{4},C_{2}=\frac{1}{6}\).  Substituting the above in (1A) gives the particular solution as\begin{align*} y_{p}\left ( t\right ) & =t\left ( \frac{3}{4}-\frac{1}{4}t+\frac{1}{6}t^{2}\right ) -\frac{1}{2}te^{-2t}\\ & =\frac{3}{4}t-\frac{1}{4}t^{2}+\frac{1}{6}t^{3}-\frac{1}{2}te^{-2t} \end{align*}

Therefore the general solution is\begin{align*} y\left ( t\right ) & =y_{h}\left ( t\right ) +y_{p}\left ( t\right ) \\ & =C_{3}e^{-2t}+C_{2}+\left ( \frac{3}{4}t-\frac{1}{4}t^{2}+\frac{1}{6}t^{3}-\frac{1}{2}te^{-2t}\right ) \end{align*}

2.4.6 Key solution for HW 3

PDF