Optimal Control Strategies for Emerging Epidemics on Heterogeneous Networks

• Activity Number: 268
• Type: Invited
• Date/Time: Tuesday, August 5, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics and the Environment
• Author(s): Krishna Pacifici*+ and Nick Meyer and Eric B. Laber and Brian Reich and John Drake
• Companies: North Carolina State University and North Carolina State University and North Carolina State University and North Carolina State University and University of Georgia
• Keywords: approximate dynamic programming ; epidemics ; infectious diseases ; invasive species

Abstract

Epidemics (e.g., invasive species, emerging infectious diseases, and computer viruses) replicating across a heterogeneous network, are responsible for a significant amount of damage each year. Deadly emerging diseases such as West Nile virus (> 1000 human deaths) and White-nose syndrome (>5 million bat deaths) represent a serious threat to ecological and environmental systems and to human health. Understanding the dynamics of these epidemics and providing mechanisms to efficiently and effectively control them is of paramount importance. We develop a spatially-explicit decision tool that provides robust optimal policies to effectively control emerging epidemics. We compare our optimal policy obtained through approximate dynamic programming and modified policy search with a model-based myopic policy and several ad-hoc policies (e.g., using graph centrality measures) on different networks. We explore different network sizes, structures (e.g., ring, star, grid) and rates of epidemic spread and dispersal. We show that our estimated policy consistently outperforms other policies and is capable of adapting to variable network structures, epidemic dynamics, and cost constraints.