多智能体系统与分布式人工智能简介

需积分: 10 132 浏览量更新于2024-08-02 收藏 793KB PDF 举报

"A Concise Introduction to Multiagent Systems and Distributed Artificial Intelligence" 是一本关于多agent系统和分布式人工智能的入门教程，由Nikos Vlassis撰写，出版于2007年。这本书由Morgan & Claypool Publishers发行，是Synthesis Lectures on Artificial Intelligence and Machine Learning系列的一部分。多agent系统（Multiagent Systems，MAS）是一种由多个自主个体（称为agent）组成的计算架构，这些个体能够通过通信和协作来解决复杂问题。在本书中，作者可能会深入探讨以下几个核心知识点： 1. **Agent的概念**：首先，会介绍什么是agent，包括其基本特征，如自主性、反应能力、适应性和社会性。自主性意味着agent能够独立做出决策，而反应能力和适应性则让agent能够对环境变化做出响应。社会性指的是agent之间可以进行交互和合作。 2. **Agent的类型**：书里可能区分了不同类型的agent，如简单反射agent、模型化agent、和社会智能agent，分别对应不同的决策层次和智能水平。 3. **Agent通信**：多agent系统中的一个重要方面是agent之间的通信协议和语言，这包括消息传递、协调策略以及如何通过共享知识和信念来实现协作。 4. **协作与冲突解决**：在多agent系统中，agent可能有各自的目标，因此如何协调行动以达到共同或相互兼容的目标是研究的关键。书中可能讨论了合作机制、谈判策略和冲突解决方法。 5. **分布式人工智能（Distributed AI, DAIS）**：DAIS是多agent系统的一个分支，它强调将AI的任务分布在多个实体之间，以利用分布式计算的优势。书里可能会介绍分布式学习、分布式优化和分布式知识表示等概念。 6. **应用领域**：多agent系统和分布式AI已广泛应用于各种领域，如智能交通、供应链管理、网络安全和游戏。书中的实例可能涵盖了这些领域的具体应用案例。 7. **进阶参考资料**：作为入门教程，这本书很可能会提供进一步阅读的参考文献，帮助读者深入学习和研究多agent系统和分布式AI的最新进展。 8. **技术框架与工具**：可能会介绍一些用于开发和模拟多agent系统的工具和框架，如Jason、JADE、Swarm和MATLAB-Simulink等。通过阅读这本书，初学者可以建立起对多agent系统和分布式人工智能的基本理解，并为进一步的学术研究或实际项目开发打下坚实的基础。

book MOBK077-Vlassis August 3, 2007 7:59

2 INTRODUCTION TO MULTIAGENT SYSTEMS

FIGURE 1.1: A robot soccer team is an example of a multiagent system

behaviors are often called heterogeneous, in contrast to homogeneous agents that are designed

in an identical way and have a priori the same capabilities. Agent heterogeneity can affect all

functional aspects of an agent from perception to decision making.

1.2.2 Environment

Agents have to deal with environments that can be either static or dynamic (change with time).

Most existing AI techniques for single agents have been developed for static environments

because these are easier to handle and allow for a more rigorous mathematical treatment. In

a MAS, the mere presence of multiple agents makes the environment appear dynamic from

the point of view of each agent. This can often be problematic, for instance in the case of

concurrently learning agents where non-stable behavior can be observed. There is also the

issue of which parts of a dynamic environment an agent should treat as other agents and which

not. We will discuss some of these issues in Chapter 7.

1.2.3 Perception

The collective information that reaches the sensors of the agents in a MAS is typically dis-

tributed: the agents may observe data that differ spatially (appear at different locations), tem-

porally (arrive at different times), or semantically (require different interpretations). The fact

that agents may observe different things makes the world partially observable to each agent,

which has various consequences in the decision making of the agents. For instance, optimal

multiagent planning under partial observability can be an intractable problem. An additional

issue is sensor fusion, that is, how the agents can optimally combine their perceptions in order

to increase their collective knowledge about the current state. In Chapter 5 we will discuss some

of the above in more detail.

book MOBK077-Vlassis August 3, 2007 7:59

INTRODUCTION 3

1.2.4 Control

Contrary to single-agent systems, the control in a MAS is typically decentralized. This means

that the decision making of each agent lies to a large extent within the agent itself. Decentralized

control is preferred over centralized control (that involves a center) for reasons of robustness

and fault-tolerance. However, not all MAS protocols can be easily distributed, as we will see

in Chapter 6. The general problem of multiagent decision making is the subject of gametheory

which we will brieﬂy cover in Chapter 3.Inacollaborative or team MAS where the agents

share the same interests, distributed decision making offers asynchronous computation and

speedups, but it also has the downside that appropriate coordination mechanisms need to be

additionally developed. Chapter 4 is devoted to the topic of multiagent coordination.

1.2.5 Knowledge

In single-agent systems we typically assume that the agent knows its own actions but not

necessarily how the world is affected by its actions. In a MAS, the levels of knowledge of

each agent about the current world state can differ substantially. For example, in a team MAS

involving two homogeneous agents, each agent may know the available action set of the other

agent, both agents may know (by communication) their current perceptions, or they can infer

the intentions of each other based on some shared prior knowledge. On the other hand, an

agent that observes an adversarial team of agents will typically be unaware of their action sets

and their current perceptions, and might also be unable to infer their plans. In general, in a

MAS each agent must also consider the knowledge of each other agent in its decision making.

In Chapter 5 we will discuss the concept of common knowledge, according to which every

agent knows a fact, every agent knows that every other agent knows this fact, and so on.

1.2.6 Communication

Interaction is often associated with some form of communication. Typically we view communi-

cation in a MAS as a two-way process, where all agents can potentially be senders and receivers

of messages. Communication can be used in several cases, for instance, for coordination among

cooperative agents or for negotiation among self-interested agents. This additionally raises

the issue of what network protocols to use in order for the exchanged information to arrive

safely and timely, and what language the agents must speak in order to understand each other

(especially, if they are heterogeneous). We will see throughout the book several examples of

multiagent protocols involving communication.

1.3 APPLICATIONS

Just as with single-agent systems in traditional AI, it is difﬁcult to anticipate the full range of

applications where MASs can be used. Some applications have already appeared, for instance

book MOBK077-Vlassis August 3, 2007 7:59

4 INTRODUCTION TO MULTIAGENT SYSTEMS

in software engineering where MAS technology has been recognized as a novel and promising

software building paradigm: a complex software system can be treated as a collection of many

small-size autonomous agents, each with its own local functionality and properties, and where

interaction among agents enforces total system integrity. Some of the beneﬁts of using MAS

technology in large systems are (Sycara, 1998):

Speedup and efﬁciency, due to the asynchronous and parallel computation.

Robustness and reliability, in the sense that the whole system can undergo a ‘graceful

degradation’ when one or more agents fail.

Scalability and ﬂexibility, since it is easy to add new agents to the system.

Cost, assuming that an agent is a low-cost unit compared to the whole system.

Development and reusability, since it is easier to develop and maintain a modular

system than a monolithic one.

A very challenging application domain for MAS technology is the Internet. Today the

Internet has developed into a highly distributed open system where heterogeneous software

agents come and go, there are no well established protocols or languages on the ‘agent level’

(higher than TCP/IP), and the structure of the network itself keeps on changing. In such an

environment, MAS technology can be used to develop agents that act on behalf of a user and

are able to negotiate with other agents in order to achieve their goals. Electronic commerce

and auctions are such examples (Cramton et al., 2006, Noriega and Sierra, 1999). One can also

think of applications where agents can be used for distributed data mining and information

retrieval (Kowalczyk and Vlassis, 2005, Symeonidis and Mitkas, 2006).

Other applications include sensor networks, where the challenge is to efﬁciently al-

locate resources and compute global quantities in a distributed fashion (Lesser et al., 2003,

Paskin et al., 2005); social sciences, where MAS technology can be used for studying in-

teractivity and other social phenomena (Conte and Dellarocas, 2001, Gilbert and Doran,

1994); robotics, where typical applications include distributed localization and decision mak-

ing (Kok et al., 2005, Roumeliotis and Bekey, 2002); artiﬁcial life and computer games, where

the challenge is to build agents that exhibit intelligent behavior (Adamatzky and Komosinski,

2005, Terzopoulos, 1999).

A recent popular application of MASs is robot soccer, where teams of real or simulated

autonomous robots play soccer against each other (Kitano et al., 1997). Robot soccer provides

a testbed where MAS algorithms can be tested, and where many real-world characteristics

are present: the domain is continuous and dynamic, the behavior of the opponents may be

difﬁcult to predict, there is uncertainty in the sensor signals, etc. A related application is robot

rescue, where teams of simulated or real robots must explore an unknown environment in

book MOBK077-Vlassis August 3, 2007 7:59

INTRODUCTION 5

order to discover victims, extinguish ﬁres, etc. Both applications are organized by the RoboCup

Federation (

www.robocup.org).

1.4 CHALLENGING ISSUES

The transition from single-agent systems to MASs offers many potential advantages but also

raises challenging issues. Some of these are:

How to decompose a problem, allocate subtasks to agents, and synthesize partial results.

How to handle the distributed perceptual information. How to enable agents to main-

tain consistent shared models of the world.

How to implement decentralized control and build efﬁcient coordination mechanisms

among agents.

How to design efﬁcient multiagent planning and learning algorithms.

How to represent knowledge. How to enable agents to reason about the actions, plans,

and knowledge of other agents.

How to enable agents to communicate. What communication languages and protocols

to use. What, when, and with whom should an agent communicate.

How to enable agents to negotiate and resolve conﬂicts.

How to enable agents to form organizational structures like teams or coalitions. How

to assign roles to agents.

How to ensure coherent and stable system behavior.

Clearly the above problems are interdependent and their solutions may affect each other.

For example, a distributed planning algorithm may require a particular coordination mechanism,

learning can be guided by the organizational structure of the agents, and so on. In the later

following chapters we will try to provide answers to some of the above questions.

1.5 NOTES AND FURTHER READING

The review articles of Sycara (1998)andStone and Veloso (2000) provide concise

and readable introductions to the ﬁeld. The books of Huhns (1987), Singh (1994),

O’Hare and Jennings (1996), Ferber (1999), Weiss (1999), Stone (2000), Yokoo (2000),

Conte and Dellarocas (2001), Xiang (2002), Wooldridge (2002), Bordini et al. (2005), Vidal

(2007), and Shoham and Leyton-Brown (2007) offer more extensive treatments, emphasizing

different AI, societal, and computational aspects of multiagent systems.

剩余83页未读，继续阅读

zangetsu

粉丝: 0
资源: 3

多智能体系统与分布式人工智能简介

multi-agent system

a concise introduction to logic 13th by Patrick J. Hurley&Lori; Watson

A Concise Introduction to Languages and Machines.pdf

A Concise Introduction to MATLAB

A Concise Introduction to Software Engineering

A Concise Introduction to Data Compression

A Concise Introduction to Data Structures using Java

A Concise Introduction to MATLAB_matlab_

A Concise Introduction to Data Compression-David Salomon

A Concise Introduction to MATLAB.zip.zip

最新资源