I’ve been studying Simon’s problem and developing simulation models using Mathematica to extend the problem to other abelian groups and hidden subgroups of order $\geq 2$. I can now obtain the $h^\perp$ to any “hidden subgroup” $h$ of the 3 abelian groups of order 8 - AbelianGroup[{2,2,2}], AbelianGroup[{2,4}], or AbelianGroup[{8}].
I am now trying to further generalize the Mathematica program in my toy examples. One tiny baby step at generalization is to change what I would call the "order" of the elements of $G$ but that word is so overloaded in this context. For example, is the identity element $e \in G$ always to be encoded as $|0 \rangle$ in the left register? What about all the other $g_i \in G$?
My question is very similar to the one posed in In the hidden subgroup problem for finite Abelian groups, where does the state $\frac{1}{\sqrt{|G|}}\sum_{g\in G} |g,0\rangle$ come from?, but I trying to make my question more pointed and the answer to my question is not found there.
When I use the the a list of $g_i$ in the sequence that Mathematica gives them to me for the assignment $g_i \rightarrow |i \rangle$ I get the correct $h^\perp$. When I manually change the sequence of $g_i$ in that list it of course remains same group of elements and I get the same cosets of $h$ in $G$, just in a different sequence, but then I don't (always) get the correct $h^\perp$ in all three toy cases.
What am I missing in the correct implementation of $\frac{1}{\sqrt{|G|}}\sum_{g\in G} | g,0\rangle$.
I hope my question is clearly conveyed.
