Consider an $ [\![n,k]\!] $ stabilizer code. Define the weight enumerator polynomial $ A(x) $ of the code as
$$
A(x):=A_0+A_1x+\dots+A_nx^n
$$
where
$$
A_j:=\frac{1}{(2^k)^2} \sum_{p \in P_n,\,\mathrm{wt}(p)=j} |\mathrm{tr}(p \Pi)|^2.
$$
Here $ \Pi $ is the projector onto the code subspace.
Is it the case that the $ A_j $ are always integers?

- 3,722
- 3
- 21
1 Answers
TL;DR: For the stabilizer codes, $A_j$ is the number of stabilizer operators with Hamming weight $j$ and thus a non-negative integer.
Suppose $g_1,\dots,g_{n-k}$ are generators of the stabilizer group $S$ of a $[\![n,k]\!]$ stabilizer code. Define $S^w:=\{s\in S\,|\,\mathrm{wt}(s)=w\}$ and $P_n^w:=\{p\in P_n\,|\,\mathrm{wt}(p)=w\}$. Then $$ \begin{align} A_j &= \frac{1}{4^k}\sum_{p\in P_n^j}[\mathrm{tr}(p\Pi)]^2\tag1\\ &= \frac{1}{4^k}\sum_{p\in P_n^j}\left[\mathrm{tr}\left(p\prod_{i=1}^{n-k}\frac{I+g_i}{2}\right)\right]^2\tag2\\ &= \frac{1}{4^k}\sum_{p\in P_n^j}\left[\frac{1}{2^{n-k}}\mathrm{tr}\left(p\sum_{b_1=0}^1\dots\sum_{b_{n-k}=0}^1g_1^{b_1}\dots g_{n-k}^{b_{n-k}}\right)\right]^2\tag3\\ &= \frac{1}{4^n}\sum_{p\in P_n^j}\left[\mathrm{tr}\left(p\sum_{s\in S}s\right)\right]^2\tag4\\ &= \frac{1}{4^n}\sum_{p\in P_n^j}\left(\sum_{s\in S}\mathrm{tr}(ps)\right)^2\tag5\\ &= \frac{1}{4^n}\left[\sum_{p\in P_n^j\setminus S}\left(\sum_{s\in S}\mathrm{tr}\left(ps\right)\right)^2+\sum_{p\in S^j}\left(\sum_{s\in S}\mathrm{tr}\left(ps\right)\right)^2\right]\tag6\\ &= \frac{1}{4^n}\left(0+\sum_{p\in S^j}\left(2^n\right)^2\right)\tag7\\ &=|S_j|\tag8 \end{align} $$ in analogy with the classical case. This gives us another way to see that $A_0+\dots+A_n=2^{n-k}=|S|$.

- 22,278
- 3
- 34
- 83
-
Do you have any thoughts on whether the $ A_j $ are always integers for non stabilizer codes? If you want I can accept this answer and make a different question specifically for non-stabilizer codes. – Ian Gershon Teixeira Feb 13 '23 at 15:23
-
This is such a nice answer I'll edit my question to specifically be about stabilizer codes and just accept this answer – Ian Gershon Teixeira Feb 14 '23 at 16:02
-
Would you happen to know a reference for this? Especially one that might contain the corresponding result "For stabilizer codes, $ B_j $ is the number of pauli operators in the normalizer $ N(S) $ with Hamming weight $ j $ and thus a non-negative integer." – Ian Gershon Teixeira Apr 21 '23 at 18:51
-
Sorry, I don't know a reference for a similar result for $B_j$ (nor for $A_j$ for that matter), but I expect a similar technique as in this answer to work for $B_j$, too. – Adam Zalcman Apr 22 '23 at 23:49