Search

In this article, the ultrafast tracking control for high-order discrete-time multiagent networks is investigated based on the predictive ability of agents. First, in view of the evolution of the original network dynamics, a distributed multistep prediction algorithm is established. Further, a control strategy containing predictive information is presented to achieve the ultrafast tracking control. Second, the problem of ultrafast tracking control with an $H_{\infty }$ performance specification is dealt with. The obtained result indicates that the robustness of the whole multiagent network can be converted into the robustness of multiple separable subsystems. Finally, the optimal ultrafast static tracking control design is provided for the single-integrator network with sampled dynamics. Two simulation cases are carried out to verify the correctness of the obtained theoretical results. [ABSTRACT FROM AUTHOR]

In modern highly automated flexible manufacturing systems, various parts are processed concurrently. Due to the concurrency and shared equipment usage, deadlock is a common problem that causes loss of productivity. When such a system is modelled by a digraph, existence of circuits in such a graph is a necessary condition for deadlock and knots and the order of circuits are closely related to impending deadlocks - a type of deadlock that is more difficult to detect. A deadlock avoidance algorithm that dynamically evaluates the order of circuits is presented. The algorithm is highly permissive since the order evaluation captures more part flow dynamics, especially when there exist multiple knots in the digraph model. It also runs in polynomial time once the set of circuits of the digraph is given. Simulation results are provided to illustrate the application of the algorithm. [ABSTRACT FROM AUTHOR]