Cellular Automata (CA) and Lattice Models represent a vital class of microscale epidemiological modeling frameworks, allowing disease spread to be examined at high spatial and individual resolution. These models discretize space into cells arranged on a lattice and time into sequential updates. Instead of relying on macroscopic averages found in classical Ordinary Differential Equation models, CA systems use explicit local interaction rules to capture heterogeneous mixing and fine-grained transmission pathways.
📐 Compartmental Structure and Flow Explanation
In CA-based epidemic models, the population is distributed across discrete spatial cells. Each agent (or aggregated subgroup) within a cell occupies a specific epidemiological state, such as Susceptible (S), Infectious (I), or Removed (R). Disease transmission and recovery are governed by rule sets that specify how an individual’s state changes according to both its own condition and the conditions of neighboring cells.
Key structural components include:
• Lattice or Cell: Represents a fixed spatial location such as a household block, community sector, or grid coordinate.
• Local Infection Flow: A susceptible proportion in cell (i, j), denoted πₛ,ᵢⱼ, becomes infected through contact with local infectious individuals πᵢ,ᵢⱼ and through interactions with infectious individuals in adjacent or connected cells.
• Iterative Progression: The system updates the entire grid synchronously (or asynchronously), applying transition rules at each discrete time step.
This rule-based approach enables detailed representation of micro-level transmission and supports the evaluation of spatially targeted interventions.
🧮 Mathematical Formulation (Discrete Update Rules)
Rather than continuous ODEs, CA models evolve through discrete-time update equations. For a simulation that tracks the proportion of infectious individuals πᵢ,ᵢⱼ within cell (i, j), transitions incorporate both local interactions and inter-cell influences weighted by connectivity coefficients ωᵢⱼ,ₚₛ:
πₛ,ᵢⱼ(t) = πₛ,ᵢⱼ(t − 1)
− β πₛ,ᵢⱼ(t − 1) πᵢ,ᵢⱼ(t − 1)
− β πₛ,ᵢⱼ(t − 1) Σ (over neighbors p, q) [ ωᵢⱼ,ₚₛ (Nₚₛ πᵢ,ₚₛ(t − 1) / Nᵢⱼ) ]
πᵢ,ᵢⱼ(t) = (1 − γ) πᵢ,ᵢⱼ(t − 1)
+ β πₛ,ᵢⱼ(t − 1) πᵢ,ᵢⱼ(t − 1)
+ β πₛ,ᵢⱼ(t − 1) Σ (over neighbors p, q) [ ωᵢⱼ,ₚₛ (Nₚₛ πᵢ,ₚₛ(t − 1) / Nᵢⱼ) ]
Here:
• V* denotes the set of neighboring cells of (i, j).
• β is the within-cell transmission coefficient.
• γ is the local recovery rate.
• Nᵢⱼ is the population in cell (i, j).
• ωᵢⱼ,ₚₛ measures the interaction strength between cells.
These equations capture both local and spatially coupled infection processes.
⚗ Parameter Definitions and Ranges
| Parameter | Interpretation | Typical Range |
|---|---|---|
| β | Local transmission probability or rate | 0.2 – 1.0 per day |
| γ | Local recovery or removal rate | 0.1 – 0.5 per day |
| Nᵢⱼ | Population size within cell (i, j) | Varies (e.g., 50 to 100,000) |
| ωᵢⱼ,ₚₛ | Inter-cell connectivity coefficient | 0.0 – 1.0 (unitless) |
| τ | Test positivity rate (if testing or quarantine is modeled) | 0.1 – 0.5 |
These parameters reflect both biological processes and spatial interaction topology.
🎯 Applicability and Limitations
| Use Case | Key Assumptions and Weaknesses |
|---|---|
| Micro-Intervention Evaluation — Effective for strategies requiring granular detail such as rapid isolation, localized lockdowns, and contact tracing at household or workplace scales. | Computational Intensity — Large lattices and high-resolution simulations demand extensive computational resources. |
| Modeling Heterogeneous Spread — Captures localized clustering, spatial mixing variability, and demographic variation across regions. | High Data Requirements — Accurate modeling requires detailed mobility and demographic data. |
| Stochastic Transmission Representation — Well suited for simulating randomness in early outbreak dynamics or small clusters. | Simplified Movement — Many lattice models use idealized random walk or grid-based movement patterns that may oversimplify real mobility behavior. |
📚 Selected References
- Kermack, W. O., & McKendrick, A. G. A contribution to the mathematical theory of epidemics.
- Ferguson, N. M., et al. Strategies for containing an emerging influenza pandemic in Southeast Asia.
- Eubank, S., et al. Modelling disease outbreaks in realistic urban social networks.
- Hethcote, H. W. The mathematics of infectious diseases.
- Tang, L., Zhou, Y., Wang, L., et al. A review of multi-compartment infectious disease models.