Benjamini et al.'s conjecture on consensus in sparse random-graph majority dynamics

Let G(n,p)G(n,p) be a random graph, let PtP_t and NtN_t denote the sets of vertices playing the two strategies at time tt, and let dd be the parameter in the condition pd/np\geq d/n. Consider the majority dynamics process with a uniformly chosen initial state. With high probability over the choice of the random graph and the initial state, the following holds. Benjamini et al.'s conjecture. If d1d\gg 1, then, for every ε>0\varepsilon>0 and all sufficiently large nn,

limtP2tN2t[(1ε)n,n].\lim_{t\rightarrow\infty}\left|\lvert P_{2t}\rvert-\lvert N_{2t}\rvert\right|\in[(1-\varepsilon)n,n].

If dd is bounded, then, for every ε>0\varepsilon>0 and all sufficiently large nn,

limtP2tN2t[(1ε/2)n,(1+ε/2)n].\lim_{t\rightarrow\infty}\left|\lvert P_{2t}\rvert-\lvert N_{2t}\rvert\right|\in[(1-\varepsilon/2)n,(1+\varepsilon/2)n].

The conjecture concerns the poorly understood evolution of majority dynamics on random graphs in the sparse regime p=o(n1/2)p=o(n^{-1/2}), beyond the regimes treated by the paper; apart from eventual periodicity, little is known about the dynamics.

Sources & referencesView supporting material

Primary source

Jordan Chellig, Calina Durbac and Nikolaos Fountoulakis, “Best response dynamics on random graphs”, arXiv:2011.12983 (2020).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.