The award will support a program of research on problems of coordination in networks. A network is a group of people or processors with links between certain pairs. Some of the problems to be studied take a network as a given and consider how to manage it efficiently. In other cases, the problem is how to design a network, as well as how to design the procedures to be followed by the agents at its nodes. In some of the design problems the agent at a node is assumed, in effect, to be a robot who behaves as the designer wishes, while in others the agent is autonomous and has independent goals that the designer has to take into account. Networks serve different purposes: transforming inputs into outputs, sending messages from one node to another, transporting commodities, coordinating mobile units, or dividing up computational tasks among processors. Part of this research considers differences between those purposes as well as their common properties. Other parts of the research pursue general themes in the design of efficient human organizations and the adaptation of market mechanisms to novel tasks of resource allocation.
