The matching process and independent process in random regular graphs and hypergraphs (Q2111785)

The paper investigates natural random greedy processes for producing matchings and independent sets in hypergraphs. In both processes, a collection (a matching or an independent set) in a hypergraph \(H\) is built by picking objects one by one. In each step, we choose a random object (an edge or a vertex, respectively) that does not conflict with previous choices; and this is repeated until no further choice is possible. The output of the algorithm is always a maximal matching or independent set of \(H\). In many cases, it is known that the aforementioned random processes build a collection (matching or independent set) that is close to being of maximum size (and not just maximal). It is of interest to estimate, for example, the number of uncovered vertices left by the matching process in particular cases of hypergraphs. The authors consider the case where the \(r\)-uniform hypergraph is randomly chosen using a degree-constrained hypergraph model. In this model, given \(0 \leq i \leq \Delta\), the number of vertices \(n_i\) of degree \(i\) is fixed, and the hypergraph is chosen at random under that constraint. The authors obtain a formula for the asymptotic number of vertices which is uncovered after the matching process, and the precise formula depends on the choice of the numbers \((n_1, \dots, n_\Delta)\). In its full generality, the formula does not give a term that is easy to calculate, but in particular cases (e.g. when every vertex has degree \(\Delta\), and \(\Delta\) is constant) it yields an explicit function in terms of \(r\) and \(\Delta\). This result is generalised to other families of hypergraphs as well, as well as for the independent process. The main tool used to analyse the random process is the differential equation method of \textit{N. C. Wormald} [Lond. Math. Soc. Lect. Note Ser. 267, 239--298 (1999; Zbl 0935.05080)].