Cooperative dynamics of loyal customers in queueing networks

We consider queueing networks （QN＇s） with feedback loops roamed by ＂intelligent＂ agents, able to select their routing on the basis of their measured waiting times at the QN nodes. This is an idealized model to discuss the dynamics of customers who stay loyal to a service supplier, provided their service time remains below a critical threshold. For these QN＇s, we show that the traffic flows may exhibit collective patterns typically encountered in multi-agent systems. In simple network topologies, the emergent cooperative behaviors manifest themselves via stable macroscopic temporal oscillations, synchronization of the queue contents and stabilization by noise phenomena. For a wide range of control parameters, the underlying presence of the law of large numbers enables us to use deterministic evolution laws to analytically characterize the cooperative evolution of our multi-agent systems. In particular, we study the case where the servers are sporadically subject, to failures altering their ordinary behavior.     

Distributed Event-Triggered Tracking Control of Heterogeneous Discrete-Time Multi-Agent Systems with Unknown Parameters

In this paper, a novel design framework is developed to solve the cooperative tracking problem of heterogeneous discrete-time multi-agent systems with unknown agent parameters and directed communication topologies. First, a distributed event-triggered observer is developed to handle the heterogeneity of the multi-agent systems. An event-triggering mechanism is proposed to reduce the amount of data transmission between neighboring agents. Then, based on the proposed observer, a discrete-time distributed model reference adaptive controller is presented to deal with the unknown parameters of the multi-agent systems. It is shown that by using the proposed observer and the model reference adaptive controller, the proposed design framework could achieve output tracking of the unknown multi-agent systems for any communication graph containing a directed spanning tree. Finally, an example is presented to illustrate the effectiveness of the proposed controller.

     

基于Multi-Agent的面向过程混杂系统的集成技术

基于 Multi-Agent提出了面向过程的混杂系统的集成框架 ,设计了用于该框架的单个 Agent的基本模型 ,分析了异构 Agent之间的关系 ,给出了一种新颖的 Agent变迁策略 .该框架能有效地刻画复杂过程控制系统的动态行为 ,便于系统的分析和设计 .     

Coordination Control for a Class of Multi-Agent Systems Under Asynchronous Switching

This paper studies the coordination control of nonlinear multi-agent systems under asynchronous switching, including consensus, tracking control, and containment. The asynchronous switching considered here means that the switching of the controller lags behind the mode's switching for each agent. So the matched controller is interrupted by the delayed switching. For the situation, the authors give some new results by applying the conventional distributed control protocol. The authors show that all agents can achieve consensus. Secondly, the authors show that all followers can track the actual leader. Thirdly, the authors show that all followers will converge to the convex hull spanned by the dynamic leaders as time goes on. Numerical simulations are also provided and the results show highly consistent with the theoretical results.     

Robust Consensus and Soft Control of Multi-Agent Systems with Noises*

This paper considers the problem of robust consensus for a basic class of multi-agent systems with bounded disturbances and with directed information flow. A necessary and sufficient condition on the robust consensus is first presented, which is then applied to the analysis, control and decision making problems in the noise environments. In particular, the authors show how a soft control technique will synchronize a group of autonomous mobile agents without changing the existing local rule of interactions, and without assuming any kind of connectivity conditions on the system trajectories.     

基于分布式协商进化算法的多Agent目标冲突消解

针对多Agent系统研究中的目标冲突消解问题,建立了在多个Agent的局部目标和系统全局目标间进行协调优化的多目标优化模型.在多Agent分布式规划的框架下,提出了一种基于遗传算法(genetic algorithm,GA)的分布式协商进化算法,用于求解多目标规划模型.针对GA搜索中保持解的多样性、提高收敛速度等问题,对选择算子进行了设计.通过仿真实验,证明新的选择算子能有效提高解的质量.最后将该算法应用于部队机动协同路线规划的目标冲突消解问题,验证了其有效性.     

Bipartite Consensus of Linear Multi-Agent Systems by Distributed Event-Triggered Control

For multi-agent systems with competitive and collaborative relationships, signed graph can more intuitively express the characteristics of their interactive networks. In this paper, the bipartite consensus is investigated for multi-agent systems with structurally balanced signed graph. In order to reduce actuation burden in dynamical network environment, the event-triggering strategy is applied to bipartite consensus protocol for the multi-agent systems. The triggered condition for each agent is designed by using its own information and transmitted information of its neighbors at sampling instant and make the number of triggers of the whole systems be reduced. Based on the distributed eventtriggered control, some sufficient conditions are derived to guarantee the leaderless and leader-following bipartite consensus. Finally, some numerical examples are shown to demonstrate the effectiveness of the theoretical results.     

Adaptive Event-Triggering Consensus for Multi-Agent Systems with Linear Time-Varying Dynamics

In this paper, the authors study the fully distributed event-triggering consensus problem for multi-agent systems with linear time-varying dynamics, where each agent is described by a linear time-varying system. An adaptive event-triggering protocol is proposed for time-varying multi-agent systems under directed graph. Based on the Gramian matrix of linear time-varying systems, the design of control gain is done and sufficient conditions ensuring the consensus of linear time-varying multiagent s...     

移动agent系统中的数据完整性分析

由于基于移动agent的分布式计算环境正日益受到越来越多的重视 ,在分析移动agent安全性问题的基础上 ,指出了通常用于保护分布式系统中安全特性方案的不足 ,提出了一个用于保证移动agent在潜在恶意主机上访问时保护收集到的数据完整性的协议。这种协议能够确保数据段的完整性与数据被收集的顺序无关 ,从而使要访问的主机可以随机被访问并能通过加密印章保证所有收集到的数据的完整性 ,因而尤其适用于比较性购物和分布式拍卖的情况。     

Event-Triggered Consensus for Discrete-Time Multi-agent Systems with Parameter Uncertainties Based on a Predictive Control Scheme

In this paper, the event-triggered consensus for linear discrete-time multi-agent systems with parameter uncertainties is investigated. The parameter uncertainty is assumed to be normbounded. An event-triggered consensus protocol based on the predictive control method is proposed to make the multi-agent system achieve consensus. And for the design of the consensus protocol, the problem of estimating the control input is transformed into the problem of estimating state differences between agents. Furthermore, the event-triggered consensus protocol proposed in this paper only demands each agent to mornitor its state to determine its event-triggered instants. A sufficient existence condition for the consensus protocol is proposed based on the linear matrix inequality. And a sufficient condition for the nonexistence of the Zeno-like behaviour is also derived. Finally, a numerical example is given to illustrate that the event-triggered consensus protocol proposed in this paper can make the multi-agent system with parameter uncertainties achieve consensus effectively.     

复杂系统、复杂网络自相似结构的涌现规律

在现实世界的复杂系统中,树状分形广泛存在.最近对于复杂网络的研究又揭示大量网络为自相似结构,随之而来的问题是:自相似结构是如何形成的.本文结合复杂系统与复杂网络,给出一个树状生长模型,通过论证揭示:生长过程及自相似结构的涌现,可集中由简单的幂律体现:幂律是自组织形成的临界状态,在它的支配下,系统得以保持有序演化发展,并涌现层层相似的自相似结构,其分形维数或相应的指数,是系统功能的度量.复杂寓于简单,简单形式的幂律蕴含了自组织演化发展并涌现自相似结构的全部过程的信息,是支配现实世界广泛现象的动力学.     

基于多Agent系统网管模型的形式化建模

运用多Agent系统实现分布、自治的网络管理是构造下一代网络管理系统的一条全新思路。然而,大多数基于多Agent系统的网管模型均采用非形式化的方法进行描述,容易造成人们对模型理解的歧义性和在系统开发过程中的随意性。一种利用有色Petri网对分布式网络管理模型进行形式化建模的方法被提出。利用该方法所建立的形式化模型,能够对分布式网管模型中多Agent系统的组成结构和运行机制进行全面刻画,能够描述模型中用到的Agent的各种特性。     

Consensus of Heterogeneous Multi-Agent Systems with Intermittent Communication

This paper studies consensus of a class of heterogeneous multi-agent systems composed of first-order and second-order agents with intermittent communication. For leaderless multi-agent systems, we propose a distributed consensus algorithm based on the intermittent information of neighboring agents. Some sufficient conditions are obtained to guarantee the consensus of heterogeneous multi-agent systems in terms of bilinear matrix inequalities(BMIs). Meanwhile, the relationship between communication duration and each control period is sought out. Moreover, the designed algorithm is extended to leader-following multi-agent systems without velocity measurements. Finally, the effectiveness of the main results is illustrated by numerical simulations.     

单边Lipschitz非线性多智能体系统一致性追踪控制

针对单边Lipschitz非线性多智能体系统,提出了一种分布式一致性控制方法.首先,构建了领导-跟随者动力学结构,用于实现单边Lipschitz多智能体系统的追踪控制.然后,设计了单边Lipschitz非线性多智能体系统的一致性控制协议,可根据智能体之间局部交互信息构建分布式反馈控制,并将系统的一致性追踪问题转化为系统...     

复杂适应性系统的仿真与研究--基于CAS理论的交通模拟

为更好地研究交通系统，以CAS理论为指导思想，利用SWARM平台建立了基于多主体离散的动态交通模拟系统。使用自底向上的建模方法进行模拟：将车辆及信号灯作为具有适应性的主体．利用元胞自动机模拟动态交通流；同时将激励学习方法与遗传算法相结合，对信号灯周期进行自适应优化。通过主体不断的“学习”，交通系统在宏观方面涌现出一定的动态特征和规律。进而，我们比较了完全自组织控制模式以及加入预警机制两种情况下各宏观量的变化情况。从实验结果可知，预警机制可以给车辆宏观方面的指导，从而提高车辆通行能力，减少交通拥塞。     

基于多智能体强化学习的多机器人协作策略研究

研究了一种基于智能体动作预测的多智能体强化学习算法. 在多智能体系统中,学习智能体选择动作不可避免地要受到其他智能体执行动作的影响,因此强化学习系统需要考虑多智能体的联合状态和联合动作.基于此,提出使用概率神经网络来预测其他智能体动作的方法,从而构成多智能体联合动作,实现了多智能体强化学习算法. 此外,研究了该方法在足球机器人协作策略学习中的应用,使多机器人系统能够通过与环境的交互学习来掌握行为策略,从而实现多机器人的分工和协作.     

Cluster-Delay Consensus for Second-Order Nonlinear Multi-agent Systems

This paper investigates the second-order nonlinear multi-agent systems subject to the cluster-delay consensus. The multi-agent systems consist of leader and agents, whose dynamics are second-order nonlinear. The objective is that the agents track the leader asymptotically with different time delays, i.e., the agents in different groups reach delay consensus, while the agents in the same group reach identical consensus. To guarantee the cluster-delay consensus for the second-order multi-agent systems, a new control protocol is proposed. Then some corresponding conditions for cluster-delay consensus are derived by using Lyapunov directed method and matrix theory. Finally, the effectiveness of the theoretical analysis results are verified by some numerical simulations.     

基于MAS的合作—竞争编队研究

基于牧羊犬跟羊群的自然现象,研究了一类具有合作—竞争拓扑的一阶多智能体系统的新型编队问题.针对每个智能体以领导—跟随者法设计了一种分布式编队控制算法,实现了合作—竞争交互网络的编队运动.利用结构平衡的独立强连通分量(SBiSCCs)推导出了一个充分条件,以确保所有智能体以分布方式进行合作—竞争编队运动.最后,通过数值仿...     

论虚拟社区的自组织现象

虚拟社区是在没有外界干预下自在、自主存在且具有组织特征的系统,它的本质是自组织的。虚拟社区的自组织表现为,具备耗散结构的一般条件,演化中呈现出自组织特征,伦理的自主与自治,以及文化形态上是自在的文化与自觉的文化的共在。这样解读虚拟社区的存在样式,对于客观、系统地认识与其相应的各种现象具有一定意义。     

A Multi-Agent Based Equipment Bidding System on the Web

This paper puts forward a new framework of multi-agent based equipment bidding system on the Web (MAEBSW). The purpose of the paper is to make the process of enterprise equipment bidding, especially the decision process of enterprise equipment bidding, more intelligent. The system discussed in this paper is an efficient method for the enterprises doing business electronically. We also present our opinion of the approaches to defining agents. We emphasize the importance of agent being able to provide explanation about its behavior for the user or other agents. Based on this opinion we design a new common architecture of most agents in MAEBSW.