An agent-based spatial temporal stochastic framework for modeling of epidemic spread and interventions – ABM-EPISPREAD

Numerous epidemic models have been implemented to study epidemic spread in time and space. Most are of compartment type [REFS].
The increase of computational speed over the last decades make it possible to implement and study sophisticated models by massive computer simulations that can then be calibrate with retrospective data, and be used to simulate scenarios of spatial spread depending on certain interventions. At the same time, geographic maps are accessible at any desired resolution including countries, regions within countries, cities and street maps within regions, with all transport lines connecting them such as streets, railway tracks and plane connections. This permits to pursue the same strategy that is becoming useful for simulations out of images in multicellular systems such as tumors, namely simulations of processes directly within the maps. Such simulations are in principle feasible down to every individual within a population like the one in France.
The objective of this proposal is to establish a pipeline permitting to integrate image information from geographical maps and traffic connections, data on population densities, and further information, in principle down to each individual human, to parameterize a spatial-temporal stochastic model of epidemic spread.
The applicants have proven expertise and publication track record on agent-based systems at many spatial resolutions in multicellular systems, which they will adapt to construct various models of epidemic spread and run them at different resolutions in space. The models will be calibrated with current pandemic data and used to predict the future development under various scenarios, including predictions of the consequences of possible political interventions.