A framework for synthesizing agent-based heterogeneous population model for epidemic simulation
Social interactions play an important role in spread of a disease. In this thesis we propose a probabilistic approach to synthesize an agent-based heterogeneous population interaction model to study the spatio-temporal dynamics of an air-born epidemic, such as influenza, in a metropolitan area. The proposed methodology is generic in nature and can generate a baseline population for the cities for which detailed population summary tables are not available. The joint probabilities of population demographics are estimated using the International Public Use Microsimulation Data (IPUMS) sample data set. Based on the population density and the socio-economic status, the population is divided into three types of residential areas. Agents, representing individuals, are assigned various activities based on their education, age, and gender. Since transportation can also influence the spread of a disease, this "activity," with a finite time span, is also assigned to individuals. The proposed approach is used for the city of Lahore, Pakistan. The agent-based model for Lahore is synthesized and a rule based disease spread model of influenza is simulated for the city population. The simulation results are visualized to analyze the spatio-temporal dynamics of an influenza epidemic for Lahore. ^
Arif Ghafoor, Purdue University.