I need to find one (or more) interesting complex power series to give to my students for their analysis exam. Ideally, this would be a power series that has interesting behavior at the boundary, i.e. does not converge everywhere/nowhere, but only at select points. To check this, they have at their disposal Abel's criterion, Dirichlet's criterion and Weierstrass' M-test. The classic examples (that they've seen) are of course the ones with coefficients $1, \frac{1}{n}$, and $\frac{1}{n^2}$. Others seem hard to find.


How about considering series such as

$$z^4 + \frac12 z^8 + \frac13 z^{12} + \dots ?$$

This would converge everywhere on the unit circle apart from the points $\pm 1$ and $\pm i$, which can easily be seen from the fact that it is obtained from the series $z + \frac12 z^2 + \frac13 z^3 + \dots$, by replacing $z$ with $z^4$, so that the single "bad point" at $z=1$ becomes the four "bad points" where $z^4=1$.

It should be easy to invent other examples along these lines where the series does not converge at a finite set of points on the circle, but the general case of looking at exactly which subsets of the circle of convergence can be realised as sets of points where a power series diverges seems to be a tough problem, as seen here on MO.


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.