In my PDE class, we are learning to solve linear homogeneous PDEs by separating variables and solving(?) for our ODEs. Working with a case with $\lambda > 0$ we are able to come to the following general solutions.

$\frac{d^2h}{dy^2}=-\lambda h \implies h=c_1 \cos{\sqrt{\lambda}y}+c_2 \sin{\sqrt{\lambda}y}$

$\frac{d^2h}{dy^2}=\lambda h \implies h=c_1 \cosh{\sqrt{\lambda}y}+c_2 \sinh{\sqrt{\lambda}y} = ce^{\sqrt{\lambda}y}$

My question is why do we assume these general solutions based off our ODE's? When does it make more sense to use the hyperbolic general solution for $\frac{d^2h}{dy^2}=\lambda h$ vs the much simpler exponential solution? Also if you could include a derivation for how we got those general solutions from the ODE's that would be fantastic.


$c_1\cosh{\sqrt{\lambda}y}+c_2 \sinh{\sqrt{\lambda}y} \neq ce^{\sqrt{\lambda}y}$

in fact :

$$\begin{align} c_1\cosh{\sqrt{\lambda}y}+c_2 \sinh{\sqrt{\lambda}y} ={c_1 \over 2}(e^{\sqrt{\lambda}y} + e^{-\sqrt{\lambda}y}) + {c_2 \over 2}(e^{\sqrt{\lambda}y} - e^{-\sqrt{\lambda}y}) = k_1 e^{\sqrt{\lambda}y} + k_2 e^{-\sqrt{\lambda}y}\\ \end{align}$$

edit : in general a linear second order ODE with constant coefficients is of the form $y'' + ay' + by + c = 0$

now something very important to keep in mind is that the set of solutions of this kind of equation is a vector space of dimension $2$ meaning if you find $2$ independent solutions you're done. because two independent solutions form a basis in a two-dimensional vector space. meaning : you can write any solution as a linear combination of them.

now how to find $2$ independant solutions ?

assume a solution is of the form $y = e^{rx}$

after plugging this in the original equation and dividing by $e^{rx}$ you get a quadratic equation which is easy to solve.

you usually either find $2$ different roots or a double root.

if you find $2$ different real roots then it's trivial that $e^{r_1x}$ and $e^{r_2x}$ are independent

if you find $2$ complex roots then one has got to be the conjugate of the other and here another very important thing you must keep in mind as well : is that if a complex function is a solution to the above equation then its real and imaginary parts are all solutions. and this is easily verifiable. and that's why in this case $\cos$ and $\sin$ appear in the solution.

there is a third case where you get only one root but you could still find a secod solution by a method called 'variation of parameters'


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.