Showing $prod_{ngeq 1} (1+q^{2n}) = 1 + sum_{ngeq 1} frac{q^{n(n+1)}}{prod_{i=1}^n (1-q^{2i})}$

I want to show

prod_{ngeq 1} (1+q^{2n}) = 1 + sum_{ngeq 1} frac{q^{n(n+1)}}{prod_{i=1}^n (1-q^{2i})}

I know one proof via self-conjugation of partition functions with Young Tableaux. But it seems not natural for me. [In the process of the proof it appears Durfee square, etc]

Is there any other (simple?) proof for this equality?

Mathematics Asked by phy_math on December 29, 2020

1 Answers

One Answer

Substituting $sqrt{q}$ for $q$, we get $$ prod_{n=1}^{infty}(1+q^n)=sum_{k=0}^{infty}{frac{q^{binom{k+1}{2}}}{prod_{i=1}^{k}{(1-q^i)}}}. $$

The left-hand side is the generating function for partitions with distinct parts (a part of each size $n$ occurs $0$ or $1$ times).

On the other hand, if a partition with distinct parts has $k$ parts, then the $i$th smallest part is of size at least $i$. Given a partition $0<lambda_1<lambda_2<dots<lambda_k$, subtract $i$ from the size of the $i$th smallest part to get a partition with parts $0lelambda_1-1lelambda_2-2ledotslelambda_k-k$ with $le k$ parts (after exluding the $0$s), whose conjugate is a partition with the largest part $le k$. The "staircase" we subtracted has $1+2+dots+k=binom{k+1}{2}$ cells. That yields the summand on the right-hand side: $$ q^{binom{k+1}{2}}prod_{i=1}^{k}{frac{1}{1-q^i}}. $$

I think this is about as simple as it gets. The Durfee square need not be involved, as you can see.

Correct answer by Alexander Burstein on December 29, 2020

Add your own answers!

Related Questions

Finding $AC$ given that $ABCD$ is inscribed in a circle

1  Asked on February 7, 2021 by questionasker


Proofing Metropolis Hastings

0  Asked on February 7, 2021 by nestroy


Integrate $2x(2x-3)^frac{1}{2} dx$.

5  Asked on February 5, 2021 by pretending-to-be-a-calculator


Rule of inference. Is the argument valid?

2  Asked on February 5, 2021 by blue


Ask a Question

Get help from others!

© 2022 All rights reserved.