Exact Solution of Granovetter's Threshold Model for a Finite Population

🕰 Orloj analysis

Tento článek odvozuje přesné analytické řešení Granovetterova prahového modelu pro kolektivní chování v konečné populaci. Zkoumá, jak individuální prahové hodnoty, parametrizované Beta distribucí, vedou k makroskopickým fázovým přechodům a analyzuje korekce konečné velikosti populace.

💡 Práce poskytuje cenné analytické řešení pro komplexní systém, které slouží jako matematický benchmark pro studium kaskád v konečných populacích, ačkoliv její přímá empirická validace je náročná.

Categories: INF-4 THE-20 INF-3 MET-1 MET-2 EMG-2 EMG-3 EMG-1 INF-5

✓ limit_reductions, modest_claims, falsifiable

⚠ Empirical validation not discussed, Assumptions about threshold distribution not empirically justified

📄 Abstract

The Granovetter threshold model formalizes collective behavior by assuming that individual agents face a binary decision to join a movement, doing so only when the number of already active participants reaches or exceeds an intrinsic, personal threshold. In this work, we derive an exact analytical expression for the probability that a cascade halts with precisely $k$ active agents in a finite population of size $N$ triggered by a single initial instigator, and use this result to obtain the scaling corrections that govern the system near its critical boundaries. By parameterizing individual threshold heterogeneity via a Beta distribution with shape parameters $α$ and $β$, we map how these micro-level predispositions aggregate into macro-level collective outcomes. Here, a small $α$ represents a high proportion of low-threshold, highly susceptible agents, while a small $β$ marks a significant density of high-threshold, conservative individuals. In the infinite-population limit, a phase transition occurs at the critical parameter $α= 1$, which separates an inactive phase from a regime of widespread mobilization. For a power threshold distribution ($β= 1$), the system exhibits a discontinuous, first-order phase transition where the active fraction jumps abruptly from 0 to 1, and the finite-size critical scaling window contracts as $N^{-1/2}$. In stark contrast, when the population features a persistent density of high-threshold agents ($β< 1$), the system undergoes an infinite-order phase transition characterized by an exceptionally smooth, continuous onset of collective activity, causing the finite-size critical region to contract at a drastically slower rate proportional to $(\ln N)^{-1}$. These analytical findings establish a mathematical benchmark for finite-size effects in behavioral cascades.

📄 arXiv abstract page 📥 PDF