arXiv++ Combinatorics

Existence of a perfect matching in a random bipartite digraph with bipartition $(V_1, V_2)$, $|V_i|=n$, is studied. The graph is generated in two rounds of random selections of a potential matching partner such that the average number of selections made by each vertex overall is below $2$. More precisely, in the first round each vertex chooses a potential mate uniformly at random, and independently of all vertices. Given a fixed integer $m$, a vertex is classified as unpopular if it has been chosen by at most $m$ vertices from the other side. Each unpopular vertex makes yet another uniform/independent selection of a potential mate. The expected number of selections made by a generic vertex $v$, i.e. its out-degree, is asymptotic to $1+\Bbb P(\text{Poisson}(1)\le m)\in (1,2)$. Aided by Matlab software, we prove that for $m=1$, whence for all $m\ge 1$, the resulting bipartite graph has a perfect matching a.a.s. (asymptotically almost surely). On the other hand, for $m=0$ a.a.s. a perfect matching does not exist, and the graph consists of a single giant component of size $2n -O(n^{1/2+o(1)})$ and possibly some components of size $O(\log n)$. This is a thorough revision of the joint paper (JCT(B) 88 (2003), 1-16) by the first author and the third author.

We introduce a random intersection graph process aimed at modeling sparse evolving affiliation networks that admit tunable (power law) degree distribution and assortativity and clustering coefficients. We show the asymptotic degree distribution and provide explicit asymptotic formulas for assortativity and clustering coefficients.