Theoretical Computer Science ( IF 1.1 ) Pub Date : 2021-04-16 , DOI: 10.1016/j.tcs.2021.04.006 Andrea Frosini , Christophe Picouleau , Simone Rinaldi
The study of the degree sequences of k-uniform hypergraphs, usually called k-sequences, has been a longstanding open problem for the case of , and the corresponding decision version was proved to be NP-complete recently in 2018 [15]. The problem can be formalized as follows: Given a non decreasing sequence of positive integers , can π be the degree sequence of a k-uniform simple hypergraph? If the answer is positive, then the sequence π is said to be k-graphic. For , that is for simple graphs, Erdös and Gallai [16] provided a characterization of the sequences that are 2-graphic (or simply, graphic). From this characterization, a polynomial time algorithm can be designed to reconstruct the incidence matrix of a graph having a given π as degree sequence (provided this graph exists). Due to the result of [15] and assuming , an efficiently computable characterization like the one for does not even exist for the case of 3-uniform hypergraphs.
Necessary or sufficient conditions for π to be k-graphic () can be found in the literature. In this paper we prove some different new conditions: first we provide sufficient and also necessary conditions for the case of k-uniform and (almost) regular hypergraphs. Then, for , we prove sufficient conditions in the case where π can be decomposed into and , and is graphic.
Most of the results are obtained by borrowing tools from discrete tomography, a current research field on discrete mathematics.
中文翻译:
一致超图的阶数序列的新充分条件
对于k-一致超图的度数序列(通常称为k-序列)的研究一直是一个长期存在的开放问题。,并且相应的决策版本最近被证明是NP-于2018年完成[15]。可以将问题形式化如下:给定正整数的非递减序列,π可以是k均匀简单超图的次数序列吗?如果答案是肯定的,则将序列π称为k图形。为了,即对于简单图形,Erdös和Gallai [16]提供了2图形(或简单图形)序列的表征。根据该特征,可以设计多项式时间算法来重构具有给定π作为度数序列的图的入射矩阵(前提是存在该图)。由于[15]的结果,并假设,一种高效的可计算特征,例如 3均匀超图的情况甚至不存在。
π为k图形的必要或充分条件()可以在文献中找到。在本文中,我们证明了一些不同的新条件:首先,我们为k一致和(几乎)正则超图的情况提供了充分必要的条件。然后,对于,我们证明了在π可以分解为的情况下的充分条件 和 , 和 是图形的。
大多数结果是通过借用离散层析成像工具获得的,离散层析成像是当前对离散数学的研究领域。