ECE Proposal Oral Exam

Title:  Mixed Centralized/Decentralized Multi-agent Decision Protocols

Committee: 

Dr. Egerstedt, Dr. Wardi, Advisor   

Dr. Ames, Chair

Dr. Shamma

Abstract: The objective of the proposed research is to develop optimization algorithms for multi-agent systems that use a mixture of centralized and decentralized elements. We propose three directions in particular. The first consists of examining how a system of agents with a central aggregator can asynchronously perform constrained optimization. The second consists of having the aggregator add noise to keep some information in the system private and analyzing the convergence of the constrained optimization algorithm with noise. Third, we propose to apply this work on a team of robots.

PhD Proposal, graduate students