Abstract:
Many applications of intelligent systems such as planning, scheduling and resource allocation can be viewed as distributed in nature. The paradigm of multiple communicating autonomous agents, known broadly as Multiagent Systems, is a good fit for such applications. However, a key outstanding challenge in Multiagent Systems is coordinating agent decisions. Distributed Constraint Reasoning (DCR) has recently emerged as a promising approach to addressing this challenge. Interdependencies between agents are modeled and reasoned about explicitly as constraints between variables assigned to different agents.

In this talk, I will discuss my recent contributions to the field of DCR. Topics I will cover include a) Asynchronous Distributed Optimization (ADOPT), the first asynchronous complete algorithm for distributed constraint optimization, b) provably correct automated problem modeling techniques, and c) heuristic solution strategies for distributed private incremental scheduling. I will discuss in particular two distributed multirobot and multiagent applications: distributed resource allocation in sensor networks and distributed meeting scheduling for personal assistant agents.

Bio:

Pragnesh Jay Modi is a Postdoctoral Fellow at Carnegie Mellon University. He holds a Ph.D. in Computer Science from the University of Southern California (2003) and a B.S. with Honors in Computer Science and Math from Carnegie Mellon University (1997). He is very active in the emerging field of distributed constraint reasoning. His dissertation entitled "Distributed Constraint Optimization for Multiagent Systems" made several influential contributions to the field. He has served as chair of the Americas School on Agents and Multiagent Systems, chair of the International Workshop on Distributed Constraint Reasoning, program committee member for the Autonomous Agents and Multiagent Systems conference and reviewer for several journals in AI including Journal of AI Research and Journal of Autonomous Agents and Multiagent Systems.