Roger Burkhart
Santa Fe Institute
The floor was opened up to discuss what the priorities should be for Swarm. Many topics had been identified for possible discussion; but, the following were emphasized most by participants:
Charles Taylor, Koji Morikawa, and John Carnahan
Biology Dept. UCLA
Chuck Taylor's group are using both Swarm and StarLogo to investigate mosquito breeding and diffusion in Africa in a project funded by the World Health Organization as part of the effort to combat Malaria. Their comparison of Swarm and StarLogo was very informative and useful. It was very easy to get simple simulations up and running with StarLogo, however they ran into some fundamental limitations with more complex models, for which they turn to Swarm.
Brian Haugh
Institute for Defense Analysis
IDA is looking at algorithms and strategies for coordinating "swarms" of ultra-small unmanned flying "drones" for military offense and defense applications.
Melissa Savage
Department of Geography
University of California, Los Angeles
Melissa is using a Swarm model built by Manor Askenazi to study the response of forests to natural and man-made perturbations, such as fires, diseases and infestations. She is using fairly abstract models of trees at the moment but she and her student Bob Bell plan to incorporate more realistic tree data and actual GIS plots of existing forests in the near future.
Primary Contact: Roger Burkhart
Primary Contact: Ted Belding
Benedikt Stefansson
Center for Computable Economics and Department of Economics
UCLA
Benedikt showed a model in which firms produce products defining points in a diversified product space, and compete to attract customers to their product line.
Gerard Weisbuch
Ecole Normale Superieure
Paris, France
Gerard discussed a number of economic models, some using a version of Swarm that we don't even remember! He also plead the case for always maintaining close association with an analytically tractable model of the system one is simulating.
J. J. Merelo
GeNeura Team
JJ Merelo reviewed his GA and Neural Net libraries for Swarm, and discussed recent upgrades.
Ted Belding
University of Michigan
Center for the Study of Complex Systems
Ann Arbor MI
Ted represented the University of Michigan Hive, describing their implementations of a number of classic social science models in Swarm, as well as their home-grown tool for managing batch runs of experimental runs, called Drone.