This question is cross-posted from math.stackexchange.com, where it did not (yet?) get any answers despite a +100 bounty.

Consider a wave equation $$\frac{\partial^2 u}{\partial t^2} = c(x)^2 \frac{\partial^2 u}{\partial x^2} \tag{1}$$ In frequency domain this becomes an ODE: $$-\omega^2 u = c(x)^2 \frac{\partial^2 u}{\partial x^2} \tag{2}$$ We can solve (2) analytically when $c(x)$ is constant, then the solution is $\exp\left(\pm\frac{i \omega}{c} x\right)$. We can also solve it analytically if $c(x)$ is piecewise constant: write in every region-of-constant-$c$ a linear combination of leftward and rightward propagating waves, impose the appropriate continuity conditions on the interfaces between regions with different values of $c$, solve for the coefficients of the linear combinations.

Suppose now that $c(x)$ is some arbitrary function. We can still write it as a limit of piecewise constant functions: $$c(x)=\lim_{\Delta\rightarrow 0^+} \sum_{i=-\infty}^{\infty} \left\{\begin{matrix}c(i\Delta) & x\in [i\Delta,(i+1)\Delta[ \\ 0 & \text{otherwise}\end{matrix}\right.$$

Let us call these piecewise constant approximations to $c$, $c_{\Delta}$ $$c_{\Delta}(x)=\sum_{i=-\infty}^{\infty} \left\{\begin{matrix}c(i\Delta) & x\in [i\Delta,(i+1)\Delta[ \\ 0 & \text{otherwise}\end{matrix}\right.$$ We can analytically solve, for any strictly positive $\Delta$, $$-\omega^2 u_{\Delta}= c_{\Delta}(x)^2 \frac{\partial^2 u_{\Delta}}{\partial x^2}$$


Is $\lim_{\Delta\rightarrow 0^+} u_{\Delta} = u$?

In other words, can we approximate general solutions of (2) by substituting in a piecewise constant approximation of $c(x)$, and then solving analytically?


Your Answer

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

Browse other questions tagged or ask your own question.