Epidemic and pandemic preparedness with rapid outbreak response rely on timely, trustworthy evidence. Mathematical models are crucial for supporting timely and reliable evidence generation for public health decision-making with models spanning approaches from compartmental and metapopulation models to detailed agent-based simulations. Yet, the accompanying software ecosystem remains fragmented across model types, spatial resolutions, and computational targets, making models harder to compare, extend, and deploy at scale. Here we present MEmilio, a modular, high-performance framework for epidemic simulation that harmonizes the specification and execution of diverse dynamic epidemiological models within a unified and harmonized architecture. MEmilio couples an efficient C++ simulation core with coherent model descriptions and a user-friendly Python interface, enabling workflows that run on laptops as well as high-performance computing systems. Standardized representations of space, demography, and mobility support straightforward adaptations in resolution and population size, facilitating systematic inter-model comparisons and ensemble studies. The framework integrates readily with established tools for uncertainty quantification and parameter inference, supporting a broad range of applications from scenario exploration to calibration. Finally, strict software-engineering practices, including extensive unit and continuous integration testing, promote robustness and minimize the risk of errors as the framework evolves. By unifying implementations across modeling paradigms, MEmilio aims to lower barriers to reuse and generalize models, enable principled comparisons of implicit assumptions, and accelerate the development of novel approaches that strengthen modeling-based outbreak preparedness.
Pandemic preparedness and rapid outbreak response are increasingly important in a globalized world with a rising frequency of epidemics and pandemics 1 . In this context, timely and reliable evidence generation is essential to support public health decision-making. Over the past decades, mathematical modeling of infectious disease dynamics has become a central tool for hypothesis generation and policy evaluation, and during the COVID-19 pandemic public institutions such as the European Centre for Disease Prevention and Control highlighted mathematical modeling as one of the principal sources of evidence for assessing the effectiveness of interventions 2 . These findings, together with the continued circulation of seasonal respiratory pathogens such as influenza and RSV, as well as re-emerging threats including Ebola and Dengue, underscore the need for sustained modeling capacity that can be rapidly mobilized to provide robust, evidence-informed support during outbreaks.
From a methodological perspective, infectious disease modeling spans a spectrum of formulations that differ in granularity, data requirements, computational cost, and interpretability. Population-based compartmental models (PBMs) describe aggregated disease dynamics using ordinary [3][4][5][6] , stochastic [6][7][8] , or integro-differential equations 9,10 . Metapopulation models (MPMs) [11][12][13] extend these approaches by incorporating spatial structure and coupling regions through mobility, thus enabling the study of spatiotemporal heterogeneity and regionally targeted nonpharmaceutical interventions (NPIs). Even further, agent-based models (ABMs) [14][15][16][17][18][19][20][21][22][23][24] resolve individuals and their interactions, allowing explicit representations of individual contact patterns, mobility, and behavior. Hybrid approaches in time and space 6,[25][26][27][28] aim to combine these paradigms in order to balance realism and computational feasibility. However, a single model formulation can barely address all facets of infectious disease mitigation and the research question typically determines which model type is most appropriate.
A wide ecosystem of open-source software supports infectious disease modeling across this spectrum, typically specializing in a particular model class and resolution (Fig. 1a). While PBMs are often reimplemented for particular applications, different packages such as epidemik 4 , MetaCast 5 , PyRoss 6 , Epydemix 7 , epipack 29 , EpiModel 30 , epidemics 31 , or EMULSION 32 also allow the flexible modeling through PBMs and MPMs; mostly using ordinary or stochastic differential equations (ODEs or SDEs). For ABMs, a variety of implementations exist. A relevant number of ABMs, such as RepastHPC 16 , OpenABM 17 , Covasim 19 , OpenCOVID 21 , EpiHiper 23 , GEMS 24 , or Vahana.jl 33 , are implemented as network-based models. These models commonly operate on coarser temporal discretizations (e.g., daily time steps) and represent locations implicitly, while mobility-or activity-based ABMs such as FRED 15 , Fig. 1: Overview of open-source, state-of-the-art infectious disease modeling frameworks, highlighting the distinctions of MEmilio. a) Overview of 21 software frameworks, depicting which model types are available through the software, which programming language has been used, if parallelization techniques are provided, which license is given, and, if unit and software tests are implemented, which code coverage is achieved. b) Statistical overview of the 10 most used programming languages in public GitHub repositories according to the search terms: SARS-CoV-2, COVID-19, influenza, Ebola, HIV, malaria, transmission, epidemiological, infection, pandemic, epidemic, endemic, pandemic spread, epidemic spread, endemic spread (accessed through the GitHub API). c) Overview of active development of MEmilio, expanding for new applications and features.
EMOD 18 , MatSim-EpiSim 20 , UHOHCoronaPolicyLab 34 , PanVADERE 35 , or the models by Goldebogen et al. 36 and Cuevas 37 explicitly simulate movement and allow explicit modeling inside locations, typically at a more fine-grained temporal resolution. Despite this breadth, the current software landscape remains fragmented across model classes, levels of resolution, and computational orientation; see Fig. 1a for an overview of 20 relevant open-source software frameworks or packages. In particular, compartmental modeling software often lacks high-performance computing (HPC) support for large-scale parameter estimation and uncertainty analysis. In addition, it offers limited flexibility for data-driven, nonexponential state transition distributions and is mostly not connected to finer-granular or hybrid modeling approaches. While several tools provide efficient simulation capabilities, integrated workflows that combine scalable calibration, uncertainty quantification, and sensitivity analysis with modern epidemiological data pipelines remain uncommon. Conversely, detai
This content is AI-processed based on open access ArXiv data.