Population protocols

Population protocols are a model of distributed computation in which
anonymous finite-state agents perform a computation by converging to a
common output value via two-way interactions.  Though the model is
simple, population protocols have a rich mathematical structure.  I
will give an overview of the model; discuss the computational power of
population protocols subject to various assumptions about which agents
can interact; and describe recent results on fast computation by
randomized population protocols, a version of the model corresponding
to well-mixed chemical solutions.

This talk describes joint work with Dana Angluin, Melody Chan, Zoë
Diamadi, David Eisenstat, Michael J. Fischer, Hong Jiang, René
Peralta, and Eric Ruppert.