基于Agent技术的智能体网格的研究

Agent能够感知外界环境的动态变化,而支持动态环境变化是网格的基本要求之一。把Agent技术引入到网格计算中,通过对Agent技术的特点和优点的分析,给出智能体网格的概念和体系结构,阐明Agent在智能体网格中,既充当Grid的使能器,又作为其功能享受者的特殊作用,最后,提出Agent技术在基于开放网格服务体系结构（OGSA）的网格计算中的应用模型。     

基于经济机制的资源管理与网格经济研究进展

本文通过回顾网格计算的特点,阐述了基于经济机制的资源分配在网格计算中的重要性和优势,分析了建立基于经济机制的网格系统所需的必要条件,通过比较有代表性的基于经济机制的资源管理系统和工作,以及参考相关的网格协议,如网格经济服务框架（GESA）等,指出了网格经济研究和实施中面临的问题与发展的方向。     

基于模拟退火遗传算法的多agent系统任务分配

针对标准的遗传算法在任务分配中收敛速度慢的问题,对多agent系统中的任务分配进行形式化描述的基础上,融合模拟退火算法的优化思想,提出了一种基于模拟退火遗传算法的任务分配方法,详细阐述了该算法的基本思想和关键步骤,并通过仿真实验进行验证。仿真实验结果表明,基于模拟退火遗传算法比标准的遗传算法具有更快的收敛速度和寻优效果。     

基于多智能体的多机群网格模型

在由多计算机机群构成的网格环境下,为了实现数据并行型计算,提出了一个基于多智能体机制的网格开发模型.给出了由多计算机机群组成的网格、逻辑计算机机群、数据并行型计算和一系列Agent的定义.利用管理智能体、独立计算智能体、协同计算智能体以及协同计算组之间的协同计算机制来实现数据并行型计算.描述了网格计算过程.实践表明,该模型有效地适应了多机群网格环境的异构性、动态性等特性,提高了计算资源的利用率.该模型适合于基于网格的并行型计算.     

竞争环境下任务分配方法的研究

以博弈论和纳什均衡理论为基础,介绍了一种基于博弈论的竞争环境下的任务分配方法,提出了竞争环境下任务的分配模型框架,引入了Agent能力的概念,并给出了竞争环境下详细的任务竞争算法和资源竞买算法,最后通过实例获得了满意的结果,同时该实例也表明了本模型的合理性和算法的有效性.     

遥感网格的数据压缩与任务分配方法研究

网格计算作为分布式计算在科学计算领域的发展方向,可以为对地观测数据的处理提供强大的计算力。在分析遥感信息服务网格节点(Remote Sensing Information Service Grid Nodes,RSSN)中网络数据传输和负载均衡两个关键问题的基础上,提出了一种有效的基于游程编码和Huffman编码的数据压缩方法和基于"计算端元"的任务分配策略,该方法针对遥感影像特点进行有效数据压缩,具有较好的压缩比,达到了17%,且能实现任务负载均衡。并在遥感信息服务网格节点计算平台上,以中国范围内1km分辨率气溶胶光学厚度(Aerosol Optical Depth,AOD)反演计算为例,从压缩率和计算时间效率方面验证和分析了上述方法的有效性。     

基于遗传算法的多智能体联盟形成机制

针对动态开放式多智能体系统中联盟的形成问题,提出“按能力分配”的联盟报酬划分规则和联盟报酬率等概念及相关命题,讨论Agent之间的协商机制和协商原则,在此基础上建立基于多智能体遗传算法的联盟形成机制。仿真计算结果表明,该联盟形成机制能够减少系统的通信量,保证所得联盟的稳定性和全局最优性,提高系统的结盟效率。     

基于网格计算的机构投资者IRMIS研究

在评述集成风险管理理论进展的基础上,指出了机构投资者集成风险管理信息系统(IRMIS)的复杂自适应系统(CAS)特征;引入网格计算思想提出了基于网格计算的机构投资者IRMIS的概念、特点及其系统原型,该原型具有多智能体(Multi-Agent)的良好结构,并基于管理熵对原型各Agent间的耦合关系进行了优化,强调风险管理战略、组织、方法、文化和过程诸要素的集成性。     

一种优化的任务分配方法研究

在计算机领域中,由于数据量和待处理任务的复杂度迅速增加,多台计算机共同协作完成任务已是未来发展的趋势,而合理的任务分配方法是高效地执行任务的前提。目前,借助多Agent系统去研究任务分配的方法是目前研究的一个热点。本文基于多Agent系统提出一种基于执行代价和传输代价的任务分配的优化方法,建立相应的任务分配模型,推导出评价任务分配优化效果的目标函数,并通过实例说明了该分配机制对多Agent系统处理问题性能的影响。结果表明,该种优化方法能有效地分析多Agent系统处理问题时的性能。     

Capability-Aware Information Aggregation in Peer-to-Peer Grids

Information aggregation is the process of summarizing information across the nodes of a distributed system. We present a hierarchical information aggregation system tailored for Peer-to-Peer Grids which typically exhibit a high degree of volatility and heterogeneity of resources. Aggregation is performed in a scalable yet efficient way by merging data along the edges of a logical self-healing tree with each inner node providing a summary view of the information delivered by the nodes of the corresponding subtree. We describe different tree management methods suitable for high-efficiency and high-scalability scenarios that take host capability and stability diversity into account to attenuate the impact of slow and/or unstable hosts. We propose an architecture covering all three phases of the aggregation process: Data gathering through a highly extensible sensing framework, data aggregation using reusable, fully isolated reduction networks, and application-sensitive data delivery using a broad range of propagation strategies. Our solution combines the advantages of approaches based on Distributed Hash Tables (DHTs) (i.e., load balancing and self-maintenance) and hierarchical approaches (i.e., respecting administrative boundaries and resource limitations). Our approach is integrated into our Peer-to-Peer Grid platform Cohesion. We substantiate its effectiveness through performance measurements and demonstrate its applicability through a graphical monitoring solution leveraging our aggregation system.     

基于群智能的P2P计算网格负载均衡算法

对等网络(P2P)计算网格是采用非集中控制的动态网络环境,在P2P网络环境的各个对等节点间均匀分配任务是网格计算的重要研究内容。传统C/S模式的负载均衡算法无法适用于分布式且动态变化的P2P网络。文章提出了一种基于群智能和多代理技术的P2P网络负载均衡算法,设计并实现了基于蚁群优化算法的分布式作业调度策略。仿真结果表明该算法是合理而有效的。     

Stabilization Parametric Region of Distributed PID Controllers for General First-Order Multi-Agent Systems With Time Delay

The stabilization problem of distributed proportional-integral-derivative (PID) controllers for general first-order multi-agent systems with time delay is investigated in the paper. The closed-loop multi-input multi-output (MIMO) framework in frequency domain is firstly introduced for the multi-agent system. Based on the matrix theory, the whole system is decoupled into several subsystems with respect to the eigenvalues of the Laplacian matrix. Considering that the eigenvalues may be complex numbers, the consensus problem of the multi-agent system is transformed into the stabilizing problem of all the subsystems with complex coefficients. For each subsystem with complex coefficients, the range of admissible proportional gains is analytically determined. Then, the stabilizing region in the space of integral gain and derivative gain for a given proportional gain value is also obtained in an analytical form. The entire stabilizing set can be determined by sweeping proportional gain in the allowable range. The proposed method is conducted for general first-order multi-agent systems under arbitrary topology including undirected and directed graph topology. Besides, the results in the paper provide the basis for the design of distributed PID controllers satisfying different performance criteria. The simulation examples are presented to check the validity of the proposed control strategy.      

Exhaustive stability analysis in a consensus system with time delay and irregular topologies

A consensus problem and its stability are studied for a group of agents with second-order dynamics and communication delays. The communication topologies are taken as irregular but always connected and undirected. The delays are assumed to be quasi-static and the same for all the interagent channels. A decentralised, PD-like control structure is proposed to create a consensus in the position and velocity of the agents. We present an interesting factorisation feature for the characteristic equation of the system which simplifies the stability analysis considerably from a prohibitively large dimensional problem to a manageable small scale. It facilitates a rare stability picture in the space of the control parameters and the delay, utilising a paradigm named cluster treatment of characteristic roots (CTCR). The influence of the individual factors on the absolute and relative stability of the system is studied. This leads to the introduction of two novel concepts: the most exigent eigenvalue, which refers to the one that defines the delay stability margin of the system, and the most critical eigenvalue, which is the one that dictates the consensus speed of the system. It is observed that the most exigent eigenvalue is not always the most critical, and this feature may be used as a design tool for the control logic. Case studies and simulations results are presented to verify these concepts.     

Designing Proportional-Integral Consensus Protocols for Second-Order Multi-Agent Systems Using Delayed and Memorized State Information

This paper is concerned with consensus of a secondorder linear time-invariant multi-agent system in the situation that there exists a communication delay among the agents in the network. A proportional-integral consensus protocol is designed by using delayed and memorized state information. Under the proportional-integral consensus protocol, the consensus problem of the multi-agent system is transformed into the problem of asymptotic stability of the corresponding linear time-invariant time-dela...     

Adaptive containment control of nonlinear multi-agent systems with non-identical agents

This paper addresses the containment control problem for a group of non-identical agents, where the dynamics of agents are supposed to be nonlinear with unknown parameters and parameterised by some functions. In controller design approach for each follower, adaptive control and Lyapunov theory are utilised as the main control strategies to guarantee the convergence of all non-identical followers to the dynamic convex hull formed by the leaders. The design of distributed adaptive controllers is based on the exchange of neighbourhood errors among the agents. For analysis of containment control problem, a new formulation has been developed using M-matrices. The validity of theoretical results are demonstrated through an example.     

数据网格环境下一种动态自适应的副本定位方法

在数据网格中，数据常常会由于性能和可用性等原因进行复制，如何有效地定位数据的一个或多个副本的物理位置是数据网格系统需要解决的重要问题，提出了一种可扩展、动态自适应的分布副本定位方法——DSRL，DSRL使用宿主结点来支持对同一数据多个副本的同时高效定位，使用本地副本定位结点来支持对副本的本地查询。DSRL提出了一种动态均衡映射方法，将全局副本定位信息均衡分布在多个宿主结点上，并且能够自适应宿主结点的动态加人或退出，详细描述了DSRL的组成，并对DSRL方法的正确性和负载平衡等特性进行了证明，分析和实验表明，DSRL方法有着良好的可扩展性、可靠性、自适应性和性能，并且实现简单，有着较好的实用性。     

Multi-parameter sensitivity analysis of power-system stability by Popov's method†

Popov's method of frequency-domain stability criteria is used to conduct the multiparameter sensitivity analysis of power system stability. The two-parameter parametric sensitivity mode] of a one-machine infinite-bus system is developed, This model involves stability measure ξ, defined in terms of Lur'o-type Liapunov functions constructed systematically. Based on the application of the model to a specific numerical example, results and conclusions are given. This type of sensitivity analysis is useful for optimum design and operation of the power system. The criterion of optimality is zero sensitivity. In general, this form of analysis is applicable to any dynamic engineering system.     

网格任务调度算法Min-min的GridSim模拟实现

该文用模拟方法研究网格中的任务调度问题．首先对Min—min算法进行分析,然后用GfidSim对Min-min调度算法进行模拟实现,阐述了实现过程,并统计模拟结果,对Min—min算法的MakeSpan和负载等性能进行了分析,验证了模拟实现过程的正确性。