Bootstrap percolation on the Hamming graphs

The \(r\)-edge bootstrap percolation on a graph is an activation process of the edges. The process starts with some initially activated edges and then, in each round, any inactive edge whose one of endpoints is incident to at least \(r\) active edges becomes activated. A set of initially activated edges leading to the activation of all edges is said to be a percolating set. Denote the minimum size of a percolating set in the \(r\)-edge bootstrap percolation process on a graph \(G\) by \(m_e(G, r)\). The importance of the \(r\)-edge bootstrap percolation relies on the fact that \(m_e(G, r)\) provides bounds on \(m(G, r)\), that is, the minimum size of a percolating set in the \(r\)-neighbor bootstrap percolation process on \(G\). In this paper, we explicitly determine \(m_e(K_n^d, r)\), where \(K_n^d\) is the Cartesian product of \(d\) copies of the complete graph on \(n\) vertices which is referred as Hamming graph. Using this, we show that \(m(K_n^d, r)=(1+o(1))\frac{d^{r-1}}{r!}\) when \(n, r\) are fixed and \(d\) goes to infinity which extends a known result on hypercubes.