Search

This paper is concerned with the distributed optimisation problem over a multi-agent network, where the objective function is described by a sum of all the local objectives of agents. The target of agents is to collectively reach an optimal solution while minimising the global objective function. Under the assumption that the information exchange among agents is depicted by a sequence of time-varying undirected graphs, a distributed optimisation algorithm with uncoordinated time-varying step-sizes is presented, which signifies that the step-sizes of agents are not always uniform per iteration. In light of some reasonable assumptions, this paper fully conducts an explicit analysis for the convergence rate of the optimisation method. A striking feature is that the algorithm has a geometric convergence rate even if the step-sizes are time-varying and uncoordinated. Simulation results on two numerical experiments in power systems show effectiveness and performance of the proposed algorithm. [ABSTRACT FROM AUTHOR]

This paper investigates the data-driven consensus tracking problem for heterogeneous multi-agent systems. All of the heterogeneous dynamics, disturbances and measurement noises are considered for the output tracking consensus of nonlinear multi-agent network under a data-driven scheme. A dynamic linearisation method is introduced to deal with the nonlinear nonaffine structures of the agents and a heterogeneous linear data model for the agent is obtained due to the heterogeneous dynamics of the agent itself. On this basis, a distributed model-free adaptive predictive control (DMFAPC) algorithm is constructed by the use of consensus errors and the robustness analysis is conducted in the presence of disturbances. Further, the results are modified by using measured outputs to replace the system outputs, and thus a measured output-based DMFAPC is presented to deal with the omnipresent measurement noises in practical applications. The two proposed DMFAPC methods are data driven since no mechanistic model information is required wherein. And thus, no unmodelled dynamics affects the consensus performance. Instead, they can improve the control performance by utilising additional predictive information. The two proposed DMFAPC methods are verified using numerical simulations. [ABSTRACT FROM AUTHOR]

In this paper, the distributed convex optimisation problem of the multi-agent system over an undirected network is investigated, in which the local objective function of each agent is only known by itself. To reduce the communication consumption between agents, a state-based dynamic event-triggered algorithm with positive minimum inter-event time (MIET) is provided, where the aperiodic information communication only occurs at some discrete triggering time instants. Moreover, the sampling control technology is combined into the previous event-triggered algorithm for verifying the event-triggered condition at every sampling time, instead of continuous access. Finally, several numerical simulations are presented for illustrating and verifying the proposed algorithms. [ABSTRACT FROM AUTHOR]

In this paper, the mechanism of the human–computer game is investigated with the help of multi-agent systems (MASs) and reinforcement learning (RL). The game is formulated as a bipartite consensus problem while the interactions among humans and computers are modelled as a multi-agent system over a coopetition network. The coopetition network associated with the multi-agent system is represented by a signed graph, where positive/negative edges denote cooperative/competitive interactions. We assume the decision mechanism of the agents are model free and each agent has to make a distributed decision by learning the input and output data from himself/itself and his/its neighbours. The individual decision is developed with the neighbours' state information and a performance index function. A policy iteration (PI) algorithm is proposed to solve the Hamilton-Jacobi-Bellman equation and obtain the optimal decision strategy. Furthermore, an actor-critic neural network is adopted to approximate the performance index and the optimal decision strategy in an online manner. The simulation results are finally given to validate the proposed reinforcement learning approach. [ABSTRACT FROM AUTHOR]

This paper focuses on the resource allocation problem(RAP) with constraints under a fixed general directed topology by using the distributed sub-gradient algorithm with event-triggered scheme in multi-agent systems, where each agent owns a cost function and its state value is bounded. The distributed sub-gradient algorithm aims to minimise the total cost by a distributed manner while achieving an optimal solution. Unlike centralised methods, the triggering condition and algorithm for each agent are fully decentralised. At each instant of time, each agent updates its state by employing the states which are collected from itself and its neighbouring agents at their last triggering time. In order to illustrate the effectiveness of the proposed sub-gradient algorithm with event-triggered control law, one simulation example is presented before the conclusion. [ABSTRACT FROM AUTHOR]

This article discusses the issue of robust adaptive neural network (NN) consensus tracking control for nonlinear strict-feedback multi-agent systems with quantised input. By combining the neural network approach with robust techniques, a novel switching function is introduced to guarantee the tracking performance of this system. To estimate the unmeasured state, an NN-based adaptive state observer is developed. Based on backstepping dynamic surface control algorithms, a robust output feedback controller is constructed to guarantee that all signals in the closed-loop system remain globally uniformly ultimately bounded. Finally, numerical simulations are carried out to demonstrate the effectiveness of the presented algorithm. [ABSTRACT FROM AUTHOR]