智能Agent建模的一种模板结构

文中讨论了设计和建造智能Ａｇｎｅｔ的一些热点问题。通过对一些基本概念和定义的评述,总结了智能Ａｇｅｎｔ应具的若干主要性质的属性。     

分布式多Agent系统流水线调度模型的矩阵分析法

采用综合逻辑的代数的方法刻画系统的动态行为,提出了分布式多Ａｇｅｎｔ系统流水线调度模型的矩阵分析法,并可直观地得到系统的Ｐｅｔｒｉ网模型,将该方法应用于自主式智能机器人流不线导航模型的分析,为系统的设计和优化提供理论依据,该方法比一般的Ｐｅｔｒｉ网分析法简捷、完备,并有严密的数学描述和分析手段。     

Event‐Based Consensus Controller for Linear Multi‐Agent Systems Over Directed Communication Topologies: A Co‐Design Approach

The paper addresses the distributed event‐triggered consensus problem in directed topologies for multi‐agent systems (MAS) with general linear dynamic agents. A co‐design approach is proposed to determine parameters of the consensus controller and its event‐triggered mechanism (ETM), simultaneously. This approach guarantees asymptotic stability along with decreasing data transmission among agents. In the proposed event‐based consensus controller, each agent broadcasts data to the neighbors only at its own triggering instants; this differs from previous studies in which continuous data streams among agents were required. Furthermore, the proposed control law is based on the piecewise constant functions of the measurement values, which are updated at triggering instants. In this case the control scheme decreases the communication network usage, energy consumption, and wear of the actuator. As a result, it facilitates distributed implementation of the proposed consensus controller for real‐world applications. A theorem is proved to outline sufficient conditions to guarantee the asymptotic stability of the closed‐loop system with the event‐based consensus controller. Another theorem is also proved to show the Zeno behavior exclusion. As a case study, the proposed event‐based controller is applied for a diving consensus problem to illustrate the effectiveness of the method.     

A REGION‐BASED MULTI‐ISSUE NEGOTIATION PROTOCOL FOR NONMONOTONIC UTILITY SPACES

Nonmonotonic utility spaces are found in multi‐issue negotiations where the preferences on the issues yield multiple local optima. These negotiations are specially challenging because of the inherent complexity of the search space and the difficulty of learning the opponent’s preferences. Most current solutions successfully address moderately complex preference scenarios, while solutions intended to operate in highly complex spaces are constrained by very specific preference structures. To overcome these problems, we propose the Region‐Based Multi‐issue Negotiation Protocol (RBNP) for bilateral automated negotiation. RBNP is built upon a nonmediated recursive bargaining mechanism which efficiently modulates a region‐based joint exploration of the solution space. We empirically show that RBNP produces outcomes close to the Pareto frontier in reasonable negotiation times, and show that it provides a significantly better performance when compared to a generic Similarity‐Based Multi‐issue Negotiation Protocol (SBNP), which has been successfully used in many negotiation models. We have paid attention to the strategic issues, proposing and evaluating several concession mechanisms, and analyzing the equilibrium conditions. Results suggest that RBNP may be used as a basis to develop negotiation mechanisms in nonmonotonic utility spaces.     

Augmented Lagrangian Pattern Search Based Multi‐Agent Model Predictive Control of Rhine‐Meuse Delta

In Large Scale Systems the concept of centrality fails due to the lack of centralized computing capability. The control of such systems has to be performed using multiple control agents. In this case, the matter of interactions among neighboring subsystems needs to be considered. In this paper, a water control system in the Netherlands is studied as a real large scale system. A multi‐agent scheme is applied to control the flow through the system which is decomposed into two interconnected subsystems. Each agent employs a model‐based predictive control (MPC) technique. The model of this large scale system is nonlinear and nonconvex. Therefore, an augmented Lagrangian pattern search optimization algorithm is used to implement multi‐agent MPC for this system. This proposed algorithm is applied by each control agent to solve its own interconnected optimization problem, at each subsystem of whole the water system. Simulation results show the effectiveness of the proposed approach.     

一种新的软件Agent设计的评价标准:用户自治性

本文提出了软件Ａｇｅｎｔ设计中的用户自治性标准,我们认为用户自治性是Ａｇｅｎｔ设计中需要考虑的重要方面,探讨了自治性标准在Ａｇｅｎｔ设计中的定义以及Ａｇｅｎｔ如何影响用户自治性,并在此基础上提出了一个新的软件系统模型。     

Agents in object‐oriented software engineering

Software engineers of multi‐agent systems (MASs) are faced with different concerns such as autonomy, adaptation, interaction, collaboration, learning, and mobility, which are essentially different from classical concerns addressed in object‐oriented software engineering. MAS developers, however, have relied mostly on object‐oriented design techniques and programming languages, such as Java. This often leads to a poor separation of MAS concerns and in turn to the production of MASs that are difficult to maintain and reuse. This paper discusses software engineering approaches for MASs, and presents a new method for integrating agents into object‐oriented software engineering from an early stage of design. The proposed approach encourages the separate handling of MAS concerns, and provides a disciplined scheme for their composition. Our proposal explores the benefits of aspect‐oriented software development for the incorporation of agents into object‐oriented systems. We also illustrate our aspect‐oriented approach through the Portalware multi‐agent system, a Web‐based environment for the development of e‐commerce portals. Copyright © 2004 John Wiley & Sons, Ltd.     

Efficient environment management for distributed simulation of large‐scale situated multi‐agent systems

Multi‐agent systems have been proven very effective for the modelling and simulation (M&S) of complex systems like those related to biology, engineering, social sciences and so forth. The intrinsic spatial character of many such systems leads to the definition of a situated agent. A situated agent owns spatial coordinates and acts and interacts with its peers in a hosting territory. In the context of parallel/distributed simulation of situated agent models, the territory represents a huge shared variable that requires careful handling. Frequent access by agents to territory information easily becomes a bottleneck degrading system performance and scalability. This paper proposes an original approach to modelling and distributed simulation of large‐scale situated multi‐agent systems. Time management is exploited for resolving conflicts and achieving data consistency while accessing the environment. The approach allows a simplification of the M&S tasks by making the modeller unaware of distribution concerns while ensuring the achievement of good scalability and performance during the distributed simulation. Practical aspects of the approach are demonstrated through some modelling examples based on Tileworld. Copyright © 2014 John Wiley & Sons, Ltd.     

On the Role of Homogeneity When Controlling Single‐Leader Networks

This paper presents an approach to controlling agent positions in single‐leader networks to target points while explicitly taking agent homogeneity into account. When the capabilities of agents to accomplish tasks at each of the targets are identical, then the label of the target points may be permuted while still expressing the same intention. In single‐leader networks which are not completely controllable, such a permutation of the target points may at times move a target closer to the system's reachable subspace, thereby allowing the network to surpass the limitations on controllability when homogeneity is not considered explicitly. To fully exploit this property in homogeneous networks, it is then necessary to find the permutation of a target point which brings it closest to the reachable subspace. However, finding this optimal permutation is shown to be in general a non‐deterministic polynomial‐time (NP)‐hard problem. Specific network topologies are identified for when finding such an optimal permutation of a target point can be advantageous when controlling single‐leader networks. Moreover, an alternate view of the problem of finding optimal permutations is presented in which clustering‐based algorithms can be applied to find suboptimal solutions.