Motivated by emerging applications in coding for molecular data storage, much attention has been paid to the intersecting set discrepancy problem, which aims to design a large family of subsets of a common labeled ground set with bounded pairwise intersection and bounded set discrepancy. In this paper, we study the maximum size of such families of $k$ -subsets with $v$ elements ground set, $t$ -bounded intersections, and zero or one discrepancy, called as balanced $(t,k,v)$ set codes. By turning this problem into a graph edge-labeling problem, we are able to determine the maximum size of codes when $k=3,4$ and $t=2,3$ for a given ground set. The constructions are based on combinatorial designs, matching decompositions and edge coloring schemes. Furthermore, we improve the upper bound for balanced $(t,k,v)$ set codes with all integers $2\leq t < k < v$ . By the powerful probabilistic argument–Kahn's Theorem, we show that the improved upper bound for any fixed integers $2 \leq t < k$ is asymptotically tight when $v$ goes to infinity.
更多
查看译文
关键词
DNA-based data storage,graph,Latin square,Howell design,matching