I had the differential equation $y'+y=\frac1x$, which I solved for $y$ as a power series:


Which was a power series at $\infty$, so it doesn't really help me much.

So my first question is whether or not $y$ is solvable here (as a power series if needed) where it actually converges.

My second question is if it pure coincidence that the summation is very similar to the power series of $e^x$.


Is there some reason for their very similar forms, or just my stumbling upon these two unrelated power series?


Multiplying both sides of the equation by $e^x$ and reversing the product rule gives $$(e^x y)' = \frac{e^x}{x} .$$

The r.h.s. famously does not admit a closed-form antiderivative, but it can be written in terms of the exponential integral function. Integrating both sides and rearranging gives (for $x > 0$) the general solution $$y = e^{-x} \left( \int_1^x \frac{e^t}{t} \,dt + C \right) .$$

One can certainly expand this in a power series (about some $x_0 > 0$), but probably it's not very illuminating.


Clear that the homogeneous solution is $y=Ae^{-x}$ so the final solution should be $$y = Ae^{-x} + f(x)$$ where $f$ is a particular solution.

Wolfram Alpha suggests the solution using special functions, so one does exist, but is not expressible in elementary terms.

  • $\begingroup$ So we can't solve for $y$ even with a power series (that converges)? $\endgroup$ – Simply Beautiful Art Apr 11 '16 at 22:38

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.