4

Let $(a_n)$ be a sequence of real numbers. The Lambert generating function of this sequence is defined as $$ L(x)=\sum_{n=1}^\infty a_n\frac{x^n}{1-x^n}. $$ I get why ordinary, exponential and Dirichlet generating functions are useful. But why are Lambert generating functions useful?

EDIT: rephrasing my question. What is an application of Lambert generating functions?

Mason
  • 4,489
Mec
  • 5,621
  • 4
    While I can't give an authoritative answer, the Wikipedia page does a decent job of suggesting some of what makes Lambert GFs interesting. In particular, if you have an ordinary GF whose coefficients is a multiplicative function, then you can do Mobius inversion to get a Lambert GF whose coefficients are also a multiplicative function (ideally a simpler one). – Semiclassical Apr 18 '25 at 22:25

1 Answers1

1

Lambert generating functions encode arithmetic information — they expose hidden recursive structures, bridge additive and multiplicative number theory, and sit at the intersection of series, products, and partitions.

A compelling application of Lambert generating functions is their ability to reveal surprising structure in arithmetic functions like the sum-of-divisors function, $\sigma_k(n)=\sum_{d|n}d^k$. For any positive integer $k$, the generating function identity $$ \sum_{n=1}^\infty \sigma_k(n)\,x^n = \sum_{n=1}^\infty \frac{n^k x^n}{1 - x^n} $$ expresses the ordinary generating function of $\sigma_k(n)$ as a Lambert series. This identity — derived by switching the order of summation in a double series — is cleanly explained here. The $k = 1$ case connects beautifully to Euler’s pentagonal number theorem, through its role in differentiating the generating function for the partition function, $\prod_{n=1}^\infty \frac{1}{1 - x^n}$. Taking a logarithmic derivative yields: $$ \frac{d}{dx} \log\left(\prod_{n=1}^\infty \frac{1}{1 - x^n}\right) = \sum_{n=1}^\infty \frac{n x^{n-1}}{1 - x^n}, $$ and multiplying by $x$ gives exactly the Lambert series for $\sigma_1(n)$. This leads to a striking recurrence involving generalized pentagonal numbers: $$ \sigma_1(n) = \sigma_1(n-1) + \sigma_1(n-2) - \sigma_1(n-5) - \sigma_1(n-7) + \cdots $$ with sign alternation and appropriate correction terms when $n$ is a pentagonal number. For a crisp explanation of this result and its connection to partitions, see Euler’s Pentagonal Number Theorem.

Mason
  • 4,489