On the number of subsequences with given sum of sequences over finite abelian \(p\)-groups (Q2478042): Difference between revisions

Let \(G\) be a finite abelian \(p\)-group, and suppose that \(S\) is a sequence of elements of \(G\). Then for \(g \in G\) and \(j \in \mathbb{N}\), denote by \(\text{N}^j_g(S)\) (respectively \(\text{N}^+_g(S)\), \(\text{N}^-_g(S)\), \(\text{N}_g(S)\)) the number of subsequences of \(S\) having sum \(g\) and length \(j\) (respectively even length, odd length, any length). The article gives some results on these numbers modulo some power of \(p\), if the sequence \(S\) is sufficiently long. Moreover, there is a similar result on the number of ways \(r_{A_1, \dots A_l}(g)\) to represent \(g\) as a sum \(\sum_{\iota=1}^l a_\iota\) (if is \(l\) sufficiently big), where the elements \(a_\iota\) come from given sets \(A_\iota \subset G\). For example suppose that \(G = C_{n_1} \oplus \dots \oplus C_{n_r}\) with \(n_1 \mid \dots \mid n_r\), and suppose that the length of \(S\) is at least \(k n_r + \sum_i (n_i - 1) + 1\). Then \(\text{N}^+_g(S) \equiv \text{N}^-_g(S) \mod p^{k+1}\). The main tool used in the proof is the group ring \(\mathbb{Z}[G]\). There are some intermediate propositions (e.g.\ about certain elements of the group ring lying in \(p^k\mathbb{Z}[G]\)), which might be of independent interest.