Percolation Models

Percolation models study connected clusters formed by open sites or edges under specified probability rules. Threshold results depend on the graph and model; finite-system spanning probabilities differ from idealized infinite-system phase transitions.

Percolation models study connectivity when sites or edges are open according to a specified probability rule. Questions include whether a connected cluster spans a finite system or whether an unbounded cluster exists in an infinite model. The graph, dependence between openings, and definition of connection determine the result; a generic count of links is insufficient.

A critical threshold in an idealized infinite model does not mean a finite system shows no change before it and complete reliable flow afterward. Cluster sizes and connection probabilities can change below the threshold, and a spanning path may be fragile. Relate the model to the actual transmission or transport process before interpreting connectivity as adoption, capacity, or benefit.

When to use it

When random availability of sites or connections affects reachability or cluster formation and the model's assumptions can be made explicit.

How it can help

Define the connectivity event and opening rule, then calculate or simulate how connectivity changes. Interpret the result together with the actual transport or transmission process rather than equating a path with successful system-wide spread.