In many societal scale engineering problems, understanding how the choices/opinions/actions of large populations or of individuals in a social network evolve is of great importance. Example application domains include transportation networks, smart power grids, public health, financial and economic networks, and social systems involving humans and AI agents. We study such systems using tools from control theory, game theory and multi-agent reinforcement learning.
Recent work:
In many coordinated planning applications, such as robot traffic management, surveillance by a team of robots and multi-vehicle routing, the problem of obtaining an optimal plan is often combinatorial or a mixed integer program. The computation time for naive optimization techniques for these problems scales exponentially, making them unsuitable for fast online repeated planning. With this motivation, we have been investigating the use of multi-agent reinforcement learning for obtaining provably safe, extremely fast, highly scalable and near-optimal policies for warehouse traffic management and team surveillance problems.
Recent work:
In many control applications, the sensing, computing and control nodes may be low powered and with limited capabilities even as the decision making is distributed, with each node having only partial information. Further, sensor and actuator signals must be communicated over a possibly a shared resource limited communication network.
Recent work: