📊 Configuration Model Agent-Based Models: Prescribing Epidemic Dynamics via Degree Distribution

The Configuration Model (CM) occupies a central position among network-based Agent-Based Models (ABMs) by explicitly encoding observed social heterogeneity into the model structure. Unlike purely random network constructions, the Configuration Model allows the modeler to prescribe a fixed degree distribution P(k), representing the exact number of contacts held by each individual agent, while connections between agents are formed randomly subject to that constraint.

This capability is essential in mathematical epidemiology because fundamental properties of disease spread—such as epidemic thresholds, outbreak probability, and the emergence of superspreading—are strongly governed by heterogeneity in contact numbers rather than by random mixing alone.

────────────────────────────────────────────

🧬 1. Compartmental Structure and Flow

The Configuration Model ABM employs standard epidemiological compartments, most commonly the Susceptible–Exposed–Infectious–Recovered (SEIR) structure, implemented in discrete time with stochastic transitions. Each agent evolves independently according to its biological state and its network connections.

Agent State Flow:

1. Susceptible (S): The infection risk of an agent is determined by its degree kₐ (number of contacts) and the infectious status of its neighboring agents, denoted by the set 𝒩(a).
2. Exposed (E): The agent enters a latent period during which it is infected but not yet infectious. Transition to the infectious state is governed by the latency rate σ.
3. Infectious (I): The agent can transmit the disease along each of its kₐ network links. Recovery occurs at a rate γ.
4. Recovered (R): The agent is removed from the transmission process and is assumed to be immune or otherwise non-infectious.

This structure ensures that agents with higher degree values are statistically more likely both to acquire infection and to transmit it onward, thereby embedding heterogeneity directly into the epidemic dynamics.

────────────────────────────────────────────

📐 2. Mathematical Formulation

Configuration Model ABMs are defined by an externally specified degree sequence drawn from a degree distribution P(k) and by probabilistic transmission rules acting along network links. The system is simulated stochastically rather than solved using global Ordinary Differential Equations.

Let 𝒩(a) denote the set of neighbors of agent a. The probability that a susceptible agent transitions from S to E during a small time interval Δt depends on exposure to infectious neighbors.

Infection Probability for Agent a:

P(S → E)ₐ = 1 − ∏ over b in 𝒩(a) of (1 − β_per × indicator that b is infectious × Δt)

Here, β_per denotes the transmission probability per contact per unit time, and the indicator function equals one when neighbor b is infectious and zero otherwise.

A defining mathematical property of the Configuration Model is that epidemic thresholds and growth conditions depend explicitly on the moments of the degree distribution, particularly the mean degree ⟨k⟩ and the second moment ⟨k²⟩. High variance in P(k) substantially lowers epidemic thresholds and increases the likelihood of large outbreaks.

Compartmental Rate Parameters Governing State Duration:

Latency rate: σ
Infectiousness (recovery) rate: γ

These rates define the expected durations of the exposed and infectious states and are typically implemented as exponentially distributed waiting times.

Parameter Definitions:

Parameter	Definition	Role in the Model
β_per	Transmission probability per single contact	Governs the likelihood of disease transmission along a network link
P(k)	Degree distribution	Specifies the probability that an agent has exactly k contacts
σ	Latent period rate (1 / T_inc)	Controls the timing of the transition from Exposed to Infectious
γ	Recovery rate (1 / T_inf)	Controls the timing of the transition from Infectious to Recovered

────────────────────────────────────────────

📊 3. Parameter Ranges (General Viral Disease)

Parameter values reflect typical biological timescales for acute respiratory viruses and empirically observed contact patterns.

Parameter	Typical Range	Unit	Context
Recovery Rate (γ)	0.07 – 0.14	day⁻¹	Infectious period of approximately 7 to 14 days
Latency Rate (σ)	0.14 – 0.25	day⁻¹	Incubation period of approximately 4 to 7 days
Transmission per Contact (β_per)	0.01 – 0.20	per contact per day	Varies by contact type, such as household versus casual contacts
Degree Distribution P(k)	Model-dependent (e.g., power-law, bimodal)	Dimensionless	Derived from empirical contact data or theoretical assumptions

────────────────────────────────────────────

🚧 4. Applicability and Limitations

Applicability – When to Use the Configuration Model:

The Configuration Model is particularly valuable when variability in contact numbers is known or hypothesized to be a dominant driver of epidemic outcomes.

1. Heterogeneity Analysis: Enables rigorous study of how non-uniform contact distributions influence epidemic thresholds, growth rates, and final outbreak size.
2. Baseline for Heterogeneous Networks: Serves as a mathematically tractable baseline that captures degree heterogeneity while remaining simpler than clustered or spatial network models.
3. Targeting High-Degree Individuals: Well suited for evaluating interventions that prioritize individuals with many contacts, such as targeted vaccination or intensified testing strategies.

Key Assumptions and Weaknesses:

1. Absence of Clustering: Configuration Models typically assume no clustering or degree–degree correlations, meaning that triangles and community structure are largely absent. This can underestimate localized transmission dynamics.
2. Static Network Assumption: The degree distribution and network connections are usually fixed, neglecting adaptive behavioral changes during an epidemic.
3. Lack of Spatial Structure: The model focuses exclusively on network topology and does not explicitly incorporate geographic space or movement unless combined with additional modeling layers.

────────────────────────────────────────────

📚 References

Anderson, R. M., & May, R. M. (1991). Infectious Diseases of Humans: Dynamics and Control.

Diekmann, O., Heesterbeek, J. A. P., & Metz, J. A. J. (1990). On the definition and computation of the basic reproduction ratio R₀ in heterogeneous populations.

Newman, M. E. J. (2002). The spread of epidemic disease on networks.

Pastor-Satorras, R., & Vespignani, A. (2001). Epidemic spreading in scale-free networks.

Wallinga, J., Teunis, P., & Kretzschmar, M. (2006). Using data on social contacts to estimate age-specific transmission parameters for respiratory-spread infectious agents.