Andrew Bernoff, Harvey Mudd College
Maria R. D'Orsogna, California State University, Northridge
Alan Lindsay, University of Notre Dame
Chad Topaz, Williams College
Alexandria Volkening, Mathematical Biosciences Institute at Ohio State
Lori Ziegelmeier, Macalester College
In the last decade, there has been a movement to describe biological and social systems via agent-based models which track individual agents (organisms, people, molecules) that interact with each other and their environment through a set of deterministic and probabilistic rules. Examples include animal swarming, traffic flow, urban crime hotspots, networks of neurons, foraging and hunting strategies, trends in social media, the antigen response in the immune system, and diffusive signaling in systems biology. While agent based models are much easier to prototype than continuum models, they offer a different set of challenges. First, the results tend to be much noisier so that identifying dynamical states entails data-analytical techniques. Second, realistic systems can often have 105 to 1010 agents or more, and many features of the system only emerge in the large population limit. This can make both simulation and subsequent data analysis challenging. Finally, these systems suffer from the curse of dimensionality familiar to mathematical biologists: the parameter spaces tend to be large and difficult to sample effectively.
The techniques that have allowed for recent progress in analyzing these models are drawn from a variety of disciplines. A goal of this workshop is to cross-train and foster the formation of collaborations between the participants so that they can formulate effective but tractable agent-based models, numerically simulate these systems efficiently (notably via parallel and cloud computing), characterize and classify the data produced by these simulations (leveraging modern ideas such as topological data analysis), and analytically describe the observed behaviors. Our hope is that this MRC will develop new connections in our community and provide our colleagues with a broad portfolio of tools to attack problems and a wider network of collaborators in this highly interdisciplinary area of research.
The application period is closed. No further applications will be accepted.
For further information, please contact the Associate Executive Director at firstname.lastname@example.org.