Automated scalable modeling for population microsimulations

The propagation of diseases within the population is an ever-reappearing hot topic in news stories. Mutations of known diseases repeatedly infected large portions of the population. In order to support the select appropriate mechanisms to help taming the spread of an epidemic, several simulation approaches have been developed to forecast the propagation behavior of diseases. Agent-based micro simulations promise to create the most detailed and accurate forecasts, but require a high modeling effort. In this paper we propose an approach to lessen this modeling effort by introducing a method that automatically creates agents for representing groups within the population based on multiple data sources (e.g. census data, vaccinations records, etc.). Our approach also facilitates combining these heterogeneous data with geographic information systems as well as dealing with incomplete data for enabling automatic and scalable creation of epidemics models for different simulation purposes in epidemiology. Two test cases were used to assess the proposition.