Intelligent Science Homepage   As My HomePage  |   As My Favorite |   Chinese Version  
 

Intelligent  Info Processing

Intelligence Science Lab
£­ Zhongzhi Shi
IntSci Research
Intelligent Systems
Intelligent Applications
Search IntSci.ac.cn

IntSci.ac.cn
 
Multi-AGent Environment - MAGE
ĦĦ

1. Introduction

2. Work flow of MAGE

3. MAGE platform architecture

4. MAGE agent architecture

5. Agent life cycle in MAGE

6. Features of MAGE

7. Application Example

ĦĦ

1. Introduction

MAGE (Multi-AGent Environment) is an agent-oriented environment for software designing, integrating and running. It provides a new computing and problem-solving paradigm in terms of agent technology. MAGE has facilities to support agent mental state representation, reasoning, negotiation, planning, cooperation and communication. It provides system designing support, description and assembling of knowledge and capability, negotiation and cooperation designing for agent-based computing on the Internet. We spent more than ten years developing MAGE system, and now MAGE has reached Version 2.0.

MAGE is a software framework fully implemented in Java language. First it simplifies the implementation of multi-agent systems through a middle-ware that claims to comply with the FIPA specifications and through a set of tools that supports the debugging and deployment phase; Second it simplifies integration of applications through multiple schemes of software reuses and an agent-oriented software design with a graphic user interface, and also it simplifies the running management through a powerful GUI with many run-time supportThe agent platform can be distributed across machines (which not even need to share the same OS) and the configuration can be controlled via a remote GUI. The configuration can be even changed at run-time by moving agents from one machine to another one when required. The system needs only Java Run Time version 1.2 or later version as a running environment.

Top of the page

2. Work flow of MAGE

According to software engineering requirements we have developed a whole procedure for software developing and system integrating in terms of MAGE, which includes requirement analyzing, system designing, agent generating and system implementing. Fig 1 demonstrates the work flow of software system developed by MAGE.

Top of the page

3. MAGE platform architecture

Fig 2 demonstrates the architecture of MAGE platform, which consists of the following components, each has specific capability set:

  • Agent Management System (AMS) exerts supervisory control over access to and use of MAGE platform. It is the managing center of MAGE platform.
  • Directory Facilitator (DF) provides yellow pages services to other agents, such as service registration, searching and updating.
  • Message Transport Service (MTS) is the default communication method between agents on different agent platforms.
  • Agent is the fundamental actor in MAGE which combines one or more service capabilities into a unified and integrated execution model. Each agent can be designed with particular function for different applications.
  • Software describes all non-agent, executable collections of instructions accessible through an agent. Non-agent software can be encapsulated with agent.

    In addition, two auxiliary modules are provided to support designing agent systems: Agent Library and Function Components. User can compose different agents by using many kinds of function components provided by MAGE. Also user can select suitable agent class from agent library.

    Top of the page

    4. MAGE agent architecture

    An agent in MAGE consists of six components: agent kernel, basic capabilities, sensor, communicator, function modules and knowledge base. Fig 3 demonstrates the detailed agent architecture.

    Top of the page

    5. Agent life cycle in MAGE

    Every agent in MAGE has a basic life cycle which includes six states: Initiated, Active, Waiting, Suspended, Transit and Unknown. And also MAGE support transition between different states, such as Invoke transform an agent from Initiated state to Active state. Refer to Fig 4 for detailed agent life cycle.

    Top of the page

    6. Features of MAGE

    6.1 Distributed computing platform

    Distributed computing allows application program elements running on different hardware platform. MAGE is a distributed computing platform, and may be distributed on different host. MAGE is built on JAVA RMI and hides all implementation details from users, thus presents a unified computing environment for users. MAGE can support building distributed application easily.

    6.2 Agent-oriented software design with a GUI

    In specifying an agent system, we have found that it is highly desirable to adopt an external viewpoint. The description of an agent system from the external viewpoint is captured in two models: Agent Model describes the hierarchical relationship among different abstract and concrete agent classes and identities the agent instances which may exist within the system their multiplicity and when they come into existence. Interaction Model describes the responsibilities of an agent class: the services it provides associated interactions and control relationships between agent classes. This includes the syntax and semantics of messages used for inter-agent communication and communication between agents and other system components: such as user interfaces. According to the two models, MAGE provides a graphic tool for agent-oriented software design: XMIDE, it is shown in Fig 5.

    6.3 Open and FIPA compliant

    The Foundation for Intelligent Physical Agents (FIPA) is an international organization that is dedicated to promoting the industry of intelligent agents by openly developing specifications supporting interoperability among agents and agent-based applications. MAGE is not isolated, it is open and FIPA compliant, so it can communicate with any agent system complying with FIPA. Agentcities is a worldwide initiative designed to help realize the commercial and research potential of agent based applications by constructing a worldwide, open network of platforms hosting diverse agent based services. The ultimate aim is to enable the dynamic, intelligent and autonomous composition of services to achieve user and business goals, thereby creating compound services to address changing needs. Up to now (2002-10-9) there are sixty agent platforms connected to agentcities.net. Our MAGE platform connected with agentcities successfully in 2002-4, and is the only agent platform connected with agentcities in China, and also become the second agent platform in Asia that connected with agentcities.

    6.4 Multiple schemes of software reuses

    Software reuse can speed software development, reduce time to market and save resources, and so MAGE provides three kinds of method to reuse software (agent software or non-agent software coded in any language):

  • Embedded
  • Executes from outside the system
  • Dynamic link library (DLL) Through these methods, MAGE can provide software reuse in different granularity:
  • Reuse of application systems
  • Reuse of sub-systems
  • Reuse of components

    Reuse of functionsMultiple methods of agent generation MAGE provides three methods of agent generation: Directly extends basic Agent class of MAGE This method aims at building new applications from MAGE. Agent Description Language (ADL) ADL is used to describe the attributes (name, address, capabilities, etc) of an agent, and then MAGE can generate an agent according to this information and this agent has the corresponding attributes. This method aims at reusing software.

    Top of the page

    7. Application Example

    MAGE is being used by a number of companies and academic groups, such as Legend , Academy of Electric Power and many others.

    Top of the page

    ĦĦ

  • About the Site | Webmaster
    Copyright © 2002-2003 Intelligent Science Research Group, at Key Lab of IIP, ICT, CAS, China.