Robust Distributed Control of Networked Systems with Linear Programming Objectives

摘要: The pervasiveness of networked systems in modern engineering problems has stimulated the recent research activity distributed control. Under this paradigm, individual agents, having access to partial information and subject real-world disturbances, locally interact achieve a common goal. Here, we consider network objectives formulated as linear program where agents' states correspond components decision vector optimization problem. To end, first contribution make is development robust continuous-time dynamics solve programs. We systematically argue that robustness properties establish for are strong can be expected programming algorithms. next design event-triggered communication protocol aforementioned algorithm. various state- based rules agents determine when they should broadcast their state, allowing us relax need continual flow between agents. Turning our attention specific control problem which algorithm applied, bargaining exchange networks. In scenario, autonomously form coalitions size two (called match) agree on how split payoff them. emphasize fair stable outcomes, whereby matched benefit equally from collaboration cannot improve allocation by unilaterally deviating outcome. synthesize algorithms converge such outcomes. Finally, focus cooperation-inducing mechanisms ensure outcome effectively realize were promised. As an illustrative example, study allocate leader role unmanned aerial vehicle (UAV) formation pairs. show strategically decide switch leading following other UAV cooperates. Throughout thesis, provably correct algorithms, making use tools controls community Lyapunov analysis Invariance Principle. Simulations optimal control, multi-agent task assignment, channel wireless networks, formations illustrate results