Equality of geometric Goppa codes and equivalence of divisors (Q1313797): Difference between revisions

The paper is concerned with algebraic-geometric codes (AG-codes). Let \(G\) and \(H\) be divisors of the same degree \(m\) on a curve of genus \(g\). \(D\) is a divisor of degree \(n\) of rational points on a curve \((D=\sum_{i=1}^{i=n} P_ i)\) and the supports of \(G\) and \(H\) are disjoint from \(D\). Let \(L(K)\) be the set of rational functions \(f\) on the curve having a divisor \(\text{div} (f)\geq -K\), \(K\) a divisor. The code \(C(D,G)\) is defined as \(C(D,G)= (f(P_ 1),\dots, f(P_ n))\mid f\in L(G)\). The main theorem of the paper is that if \(2g-1<m< n-1\) then \(C(D,G)= C(D,H)\) if and only if \(G\) and \(H\) are equivalent with respect to \(D\) (i.e. \(G= H+\text{div} (u)\) for some rational function \(u\) satisfying \(u(P_ i)=1\), for all \(i=1,\dots, n)\). For \(m=2g-1\) or \(m=n-1\) there are some special cases in which \(C(D,G)= C(D,H)\) while \(G\) and \(H\) are not equivalent themselves (but are equivalent up to a point). These cases are listed in Theorem 4.14, b, c, d and e. The main theorem leads in a natural way to a characterization of (formally) self-dual AG-codes \((n>2g+2)\): \(C(D,G)\) is (formally) self- dual if and only if there exists a differential form \(\eta\) with single poles at the \(P_ i\) such that \(2G=D+ \text{div}(\eta)\) (in case of self duality \(\text{res}_{P_ i} (\eta)=1\)). The paper concludes with counting the number of AG-codes on a curve arising from a divisor \(D\) in terms of the generalized zeta-function of the curve with respect to \(D\).