0

So I saw this differential equation $$x^2y^{\prime \prime} - xy^{\prime} + y = \ln(x)$$ let $x=e^t$ or $t = \ln x$ $$\frac{dy}{dx} =\frac{dy}{dt}\frac{dt}{dx}= \frac{1}{x}\frac{dy}{dt}$$ $$\frac{d^2y}{dx^2} = \frac{1}{x}\frac{d}{dx}\left(\frac{dy}{dt}\right) + \frac{dy}{dt}\left(-\frac{1}{x^2}\right)$$ $$\frac{d^2y}{dx^2} = \frac{1}{x}\left(\frac{d^2y}{dt^2}\frac{1}{x}\right) + \frac{dy}{dt}\left(-\frac{1}{x^2}\right)$$

so here how did we derive $\frac{d}{dx}(\frac{dy}{dt}) = \frac{1}{x}\frac{d^2y}{dt^2}$, even the textbook I used this example from has no derivation in it

2 Answers2

4

This can easily be resolved by not using this terrible mess of notation which is unfortunately very common. So what is actually going on is we are considering two functions, namely $y$, which is the solution to the ODE, and $t$, defined by

$$t(x)=\ln x.$$

When we then write "$y(t)$", what we really mean is the function $y\circ t$. This is what's going on when we make a change of variables: we compose our function with another function. What the equation you're confused about is really saying then is that

$$(y'\circ t)'(x)=\frac{1}{x}(y''\circ t)(x)$$

(you could alternatively write $y''(t(x))$ instead of $(y''\circ t)(x)$ if that makes it clearer for you). But this them becomes a really easy to see consequence of the chain rule, which tells us that

$$(y'\circ t)'=t'\cdot(y''\circ t),$$

and so as

$$t'(x)=\frac{1}{x},$$

we get the result above.

Lorago
  • 9,239
  • 1
    This is going to be highly subjective, but frankly, I find your (i.e. Lagrange's) notation to be much more of a "terrible mess" than the Leibniz notation when it comes to understanding what's going on with chain rule in single variable calculus. The visualisation of "fraction-like" terms with "cancellation" is just simpler to me. – Deepak Feb 03 '23 at 16:00
  • 1
    @Deepak of course this is highly subjective, and if it makes it clearer for you then that's great. But in my opinion Leibniz's notation really obscures what's going on most of the time, and stating something like the chain rule as $\frac{\mathrm{d}y}{\mathrm{d}x}=\frac{\mathrm{d}y}{\mathrm{d}t}\frac{\mathrm{d}t}{\mathrm{d}x}$ really hides what is actually going on imo (largely because it's not at all clear what functions are being evaluated where, and also the variable names are not an inherent properly of the functions). – Lorago Feb 03 '23 at 16:05
  • I think this answer does a great job at getting to the heart of the issue – Lorago Feb 03 '23 at 16:06
  • 1
    @Lorago I certainly agree, it especially doesn't help that most people who sing the praises of Leibniz notation hardly understand what a differential really is. – Philosophiæ Feb 03 '23 at 16:09
2

Chain rule:

$$\frac{d\frac{dy}{dt}}{dx} = \frac{d\frac{dy}{dt}}{dt} \cdot \frac{dt}{dx} = \frac1x \frac{d^2y}{dt^2}$$

user170231
  • 19,334