The size of the giant joint component in a binomial random double graph (Q2227829)

This paper studies the size of the giant joint component in a binomial random double graph. A double graph is formed by superposing two graphs \(G_1=(V, E_1)\) and \(G_2=(V,E_2)\) over the same vertex set \(V\). The binomial random double graph \(G(n,p_1,p_2)\) is a graph where \(G_1\) and \(G_2\) are independent binomial random graph over \([n]\) and the edge probabilities are \(p_1\) and \(p_2\). For \(p_1=\lambda_1/n\) and \(p_2=\lambda_2/n\), let the solution of the equation \(\beta=\mathbb{P}(Po(\lambda_1\beta)>0)\mathbb{P}(Po(\lambda_2\beta)>0)\) be denoted by \(\beta(\lambda_1,\lambda_2)\). Let \(C=\partial\{(\lambda_1,\lambda_2):\beta(\lambda_1,\lambda_2)=0\}\). For \((\lambda_1,\lambda_2)\in\mathbb{R}^+\backslash C\) the number of vertices in the giant joint component of \(G(n,\lambda_1/n,\lambda_2/n)\) is \(\beta(\lambda_1,\lambda_2)n+o_p(n)\) as \(n\) goes to infinity. Moreover, it is shown that when \(\beta(\lambda_1,\lambda_2)\) is positive, the giant joint component is unique.