Probabilistic graph models in which vertices are placed in a geometric space (often at random) and edges are formed according to geometric proximity or a distance-dependent rule, providing a bridge between continuum geometry and random graph theory.

Principle

Embed vertices in a metric space and connect pairs based on a threshold distance or a kernel of connection probability; study graph properties (connectivity, component sizes, clique numbers, percolation) as functions of the underlying geometry, point process, and scaling regime.

Demonstration

In the Gilbert (unit‑disk) model on the plane, n points sampled uniformly in a square are connected when their Euclidean distance is below radius r; as n grows and r scales appropriately, one observes phase transitions in connectivity and emergence of a giant component.

Misapplication

Treating random geometric graphs as Erdős–Rényi graphs without accounting for spatial correlation or assuming independence of edges leads to incorrect thresholds and false conclusions about clustering and path lengths.

Consequence

Random geometric graphs model spatial networks (wireless, sensor, ecological) with realistic locality, provide asymptotic laws for connectivity and spectral properties, and inform algorithms that exploit geometry for routing and clustering.

Reversal

Reverse by ignoring geometry and using purely combinatorial random graphs; this simplifies analysis but omits spatial constraints, locality effects, and geometry-driven phenomena like percolation thresholds dependent on dimension.

Boundary

Framework depends on the ambient metric space, point process (Poisson, binomial), and connection rule; many results assume independence, stationarity, or homogeneity that fail in heterogeneous or constrained environments.

Semantic Tension

Tension lies between geometric/dependent-edge models and classical independent-edge random graphs: geometry induces dependencies and locality that conflict with mean-field approximations common in combinatorial theory.

Synthesis

Random geometric graphs integrate stochastic geometry and graph theory by placing nodes in space and forming edges by proximity rules, capturing spatial dependence and enabling analysis of connectivity, scaling limits, and applications in networked systems.