Decentralized Control of Complex Systems: Managing Uncertainties and Multi-objective Optimization with Multiple Controllers

This dissertation explores how learning-based adaptive controllers can be designed to operate in environments with other controllers, meeting multiple objectives while remaining robust to uncertainties and disturbances. The central hypothesis is that optimal control for large, complex systems can be broken down into a series of decentralized optimization problems, solved independently by each controller at each time step. Under appropriate assumptions, these controllers can learn solutions that ensure system safety, optimality, and resilience to disturbances.