Synchronization Analysis And Control Based On A Power-Driven Moving Agent Network

In the past decade, network-based approaches have attracted an increasing interest and have been proved to be prominent candidates to investigate the collective dynamics in many branches of science and engineering. Various complex systems can be described as complex networks, where the individual units are abstracted by nodes, and interactions be-tween individuals are represented as edges. Thus, dynamical evolution of these systems can be investigated under the architecture of networks. As a ubiquitous process in many dif-ferent complex systems, synchronization on networks has been extensively and intensively studied. The key point is to uncover the essential factors that determine the synchronization process on networks. Furthermore, we can control the occurrence of synchronization, en-hancing it when favorable and suppressing it when harmful. In this dissertation, we propose two power-driven mobile agent networks to represent real world systems, and analyze the synchronization process on them. Moreover, adaptive rules and pinning control strategies are introduced to control synchronization of the power-driven mobile agent network. The main contents are outlined as follows:1. A general power-driven mobile agent network model is proposed as a simple repre-sentation of real-world systems in which couplings between individuals are relevant to energy. To investigate the effect of power distribution, synchronization is further explored for the power-driven mobile agent network. Under the constraint of fast-switching, we theoretically and numerically show that synchronization of the agent network is determined by the power density which is independent of both the power distribution and the size of network.2. A blinking power-driven mobile agent network model is proposed as a simple rep-resentation of real-world systems in which influence ability of individuals exhibits an intermittency feature. Through theoretical analysis and numerical simulations, we discuss the synchronization issue of the network in two cases:random blinking and ordered blinking, and show that synchronizability of the presented network is enhanced dramatically under ordered blinking case. In addition, effect of blinking behaviors including blinking period and duty ratio, on network synchronization is investigated by evaluating the convergence time.3. Synchronization issue of a dynamical network with general switching topology is investigated. Compared to networks with static topology and fast-switching topol-ogy, it is rather difficult to analyze the synchronization issue of a dynamical network with a general switching topology. By constructing a common Lyapunov function, we show that local and global synchronization for a linearly coupled network with switching topology can be evaluated by the time average of second eigenvalues cor-responding to the Laplacians of switching topology. Several simulations illustrate the effectiveness of the obtained results.4. Synchronization control problem of power-driven mobile agent network is investi-gated. For the general power-driven mobile agent network, we design a power adap-tive law to guarantee its synchronous motion. For the blinking power-driven mobile agent network, we design a blinking adaptation strategy to enhance its synchroniz-ability. Moreover, we studied the problem of synchronizing the general power-driven mobile agent network by means of pinning control. Two pinning strategies are used when pinning the network, and the corresponding control performances are com-pared.The conclusions and perspectives are presented in the end of the dissertation.