Codes over rings of size \(p^2\) and lattices over imaginary quadratic fields (Q964414): Difference between revisions

This paper explores the connection between linear codes over finite fields (and even certain finite rings) and lattices over the ring of integers of a number field. For this paper, the number field is an imaginary quadratic field \(K={\mathbb{Q}}(\sqrt{-\ell})\), where \(\ell\) is a square-free positive integer with \(\ell \equiv 3\pmod 4\). If \(\mathcal{O}\) is the ring of integers of \(K\) and \(p\) is a prime, \(p\nmid \ell\), then let \(R={\mathcal{O}}/p{\mathcal{O}}\). Suppose \(\Lambda\) is a lattice over \({\mathcal{O}}\) corresponding to a linear code \(C\) over \(R\). To emphasize the dependence of \(\Lambda\) on \(\ell\), we call \(\ell\) the ``level'' of \(\Lambda\) (or of \(C\)). The theta series of \(\Lambda\) can be expressed in terms of ordinary theta functions and the complete weight enumerator \(\text{cwe}_C\) of \(C\) (and, of course, \(p\) and \(\ell\)). The authors ask two natural questions. First, how do the theta functions of the same code differ for different levels \(\ell\)? In this direction, the following fact is proven: If \(\ell,\ell'\) are levels as above with \(\ell'<\ell\), if \(\Lambda\) has level \(\ell\) and \(\Lambda'\) has level \(\ell'\), if \(\theta\) is the theta series of \(\Lambda\), and if \(\theta'\) is the theta series of \(\Lambda'\), then the first \((\ell+1)/4\) terms in the \(q\)-series expansion of these theta functions agree. (This was actually proven by Shaska-Wijesiri in an earlier paper.) Second, can inequivalent codes give rise to the same theta functions for all levels \(\ell\)? The authors conjecture that if \(C\) and \(C'\) have the same same theta functions for all sufficiently large levels \(\ell\) then \(C\) and \(C'\) have the same symmetric weight enumerator. (Actually, they give precise, and much more technical, conjectures in the paper. This is a simplified version of them.) The authors verify this conjecture by explicit computations (mostly omitted) for \(p=2,3,5\).