The Kendrick modelling platform: language abstractions and tools for epidemiology

Area of Science:

• Epidemiology
• Computational Biology
• Mathematical Modeling

Background:

• Mathematical and computational models are crucial for studying infectious diseases but are often complex and difficult to reproduce.
• Intertwined epidemiological concerns and low-level implementation details hinder accessibility for non-computing scientists.
• Existing models lack ease of definition, reuse, and reproduction.

Purpose of the Study:

• To simplify the definition, reuse, and reproduction of compartmental epidemiological models.
• To facilitate the implementation of diverse simulation approaches with minimal programming knowledge.
• To bridge the gap between epidemiological research and computational implementation.

Main Methods:

• Development of Kendrick, a domain-specific language (DSL).
• Representation of epidemiological concerns as stochastic automata.
• Combination of automata using tensor-algebra operators for model construction.

Main Results:

• Kendrick supports a wide range of epidemiological concerns (multi-species, spatial, policy, demographics) independently.
• Models can be defined, modified, and reused modularly without low-level implementation details.
• Implementation of epidemiological models requires significantly reduced programming expertise.

Conclusions:

• Kendrick is a unique DSL for epidemiological modeling that abstracts implementation complexities.
• It promotes model reproducibility and reusability through a clear separation of concerns.
• Future work aims to expand Kendrick's capabilities to non-compartmental models and enhance tool interoperability.