Explosive Percolation

Explosive percolation is a class of dynamical percolation models in which a giant connected component emerges through a discontinuous, first-order phase transition rather than the continuous second-order transition predicted by classical percolation theory. In explosive percolation models, edges are added according to competitive rules — for example, choosing the edge that minimizes the product of the component sizes it would connect — that actively suppress the growth of large components until a critical moment, at which point the system collapses into a spanning cluster in a single macroscopic step.

The phenomenon was first reported by Achlioptas, D'Souza, and Spencer in 2009 and was initially controversial: some researchers argued that the apparent discontinuity was a finite-size effect that would vanish in the thermodynamic limit. Subsequent analysis confirmed that explosive percolation is a genuine first-order transition in certain competitive growth models, though the transition is 'weakly' first-order with unusual scaling properties that differ from both standard first-order and second-order transitions.

Explosive percolation matters because it shows that the order of the percolation transition is not universal. It depends on the growth rules. In real systems — such as financial networks where institutions deliberately avoid overconnectedness, or social networks where homophily slows cross-community linking — the suppression of large-component growth can create conditions for explosive rather than continuous percolation. The policy implication is that gradual stress accumulation can mask a system that is closer to catastrophic failure than continuous models predict.