Development of Multi-Agent System Based on ...

106 downloads 0 Views 66KB Size Report
Keyword: Multi-Agent System, FIPA Specification, JADE, JACK Intelligent Agents. 1. INTRODUCTION: Today agent based technology is one of the emerging ...
International Conference on Reliability, Infocom Technology and Optimization (ICRITO'2010)

Development of Multi-Agent System Based on Frameworks Dr. Mohammad Ubaidullah Bokhari*, Sadaf Ahmad**, Faheem Syeed Masoodi** *Chairman, Department of Computer Science, AMU, Aligarh **Research Scholar, Department of Computer Science, AMU, Aligarh

ABSTRACT: Multi-agent system is one of the emerging trends being used today. The application of multi-agent systems is spreading in almost every field. It can provide new solutions to very complex and restrictive system. The term “Agent” though known from last few decades, yet it still is a very popular technology and is being used radically for the development of distributed software systems. This paper mainly focuses on the different framework available for the development of Multi-Agent system. Keyword: Multi-Agent System, FIPA Specification, JADE, JACK Intelligent Agents 1. INTRODUCTION: Today agent based technology is one of the emerging technologies in distributed Environment. It is widely used because it helps us to design complex system with less effort. Agents are the software programs, having the ability to take their own decision according to the requirement. It can be move from one system to other system and have the ability to learn new information about its environment so that they can make some improvement on itself. When these intelligent agents work together to achieve a same goal then the overall system is known as Multi-Agent System (MAS). The key principle of multi-agent systems is that each agent pursues its own local goal and that the global goal comes out from the interaction of agents[2]. It is one of the rising technologies and to develop a multiagent based distributed software system lots of frameworks are available like Jade, FIPA-OS, JACK Intelligent Agent, JIAC and etc. Two of the frameworks discussed in this paper are JADE and JACK Intelligent Agents. These environments provide some predefined agent models and tools which help developers in the system development and reduce the amount of effort required in designing a multi agent system. However, agent-based technologies still need to be mature so that more work can be done to improve the quality of the system [1].This paper is distributed in five sections. In Section second, we will discuss the Multi agent system and the third section will describe FIPA Specification and then the fourth section will present JADE (Java Agent Development Framework), section five describe the JACK and finally the conclusion. 2. MULTI-AGENT SYSTEM: A multi-agent system (MAS) is a system in which multiple autonomous agents work together to achieve a same goal by using its own local information see fig 1.

International Conference on Reliability, Infocom Technology and Optimization (ICRITO'2010)

Fig 1: Multi-Agent System Agent-based systems created excitement among Artificial Intelligent Communities about the new paradigms that it offers for conceptualizing, designing and implementing systems [2]. These systems may consist of similar or different types of agents. Agents fall in two categories either benevolent or self-interested. Cooperative agents work together to achieve a set of shared goals, whereas selfinterested agents have their separate goals but they still interact to attain their own goal. Certain reasons of interaction between agents in a multi-agent system are: 1. Either to share the resources 2. Possible relation between Sub-problems Autonomous agent should direct contract to the less autonomous characteristics of agents in usual distributed processing applications. Many issues arise, in the perspective of design, implementation, and assessment. The very common problem that researchers face in dealing with these issues is the possibility that the information an agent is using to make its decisions is incomplete, out-of-date or inconsistent with that of other agents [3]. 3. FIPA SPECIFICATIONS The Foundation for Intelligent Physical Agents (FIPA) is a non-profit association registered in Geneva, Switzerland. FIPA’s purpose is to promote the success of emerging agent-based applications, services and equipments. This goal is pursued by making available the internationally agreed specifications that maximize interoperability across agents in a timely manner [4]. The Table below shows the two different specifications of FIPA and their parts. FIPA-Specification FIPA-97 Specification Normative Part1 Part2 Part3 Part4 Part5 Part6 Part7

Informative

Agent Management Agent Communication Language(ACC) Agent Software Integration Personal Travel Assistance Personal Assistant Audio-Visual Entertainment and Broadcasting Network Managenet and Provisioning

International Conference on Reliability, Infocom Technology and Optimization (ICRITO'2010)

FIPA-98 Specification Normative Part-8 Part-9

Human/Agent Interaction

Part-10 Part-11 Part-12

Agent Security Agent Mobility Ontology Service information application specification

Informative

Product Design and Manufacturing

FIPA-97 Developers Guide Table-1: FIPA Specification FIPA-97 Specification is the 1st output document of FIPA which comprises of seven parts, out of which first three are normative type which mainly focuses on technical aspects of multi-agent systems. These three agents are Agent Management System (AMS), Agent Communication Channel (ACC) and agent/software interaction. The next four are known as informative type which would explain how to use AMS, ACC, Agent/Software Integration to implement the four selected applications which are Personal Travel Assistance, Personal Assistant, Audio-Visual Entertainment and Broadcasting, Network Management and Provisioning. [5, 6,] In the year 1998 FIPA produce its second version of FIPA 97 that is FIPA 98. FIPA 98 describes six – part specification: the normative specifications are Human/Agent Interaction, Agent Security, Agent Mobility, and Ontology Service; informative application specification other two are informative specifications Product Design and Manufacturing and FIPA 97 Developers’ Guide. Agents use Agent Communication Language (ACL) to exchange messages [1]. Part-13

4. JADE JAVA AGENT DEVELOPMENT FRAMEWORK: JADE (Java Agent Development Framework) is a software framework to make easy the development of multi-agent applications in compliance with the FIPA specifications. Through JADE the development of multi-agent system becomes easier. JADE is implemented in Java language. JADE uses an agent model and Java implementation that allow good runtime efficiency, software reuse, agent mobility and the realization of different agent architectures [1, 8, 7, 9]. JADE creates containers for agents, each of which can either be on same computing system or in different systems. There is a direct connection between containers these containers collectively form a platform. There must be a main container in each platform which itself contain two specialized agents called the AMS (Agent Management System) which is the authority in the platform to kill other agents, kill containers, and shut down the platform and the DF (Directory Facilitator) which provide a yellow pages service to platform [10].

International Conference on Reliability, Infocom Technology and Optimization (ICRITO'2010)

Fig -2 JADE

JADE design gives programmers an easy access of FIPA standard assets. The scalability and fault tolerance are the two very important issues for distributed robust software infrastructures. When discussing scalability, it is necessary to first state with respect to which variable; in a Multi Agent System, the three most interesting JADE agents are very scalable with respect to the number of simultaneous conversations a single agent can participate in. when an agent engages in a new conversation, no new threads are spawned and no new connections are set up, just a new behavior object is created [11]. So the only overhead associated to starting conversations is the behavior object creation time and its memory occupation; agents particularly. 5. JACK INTELLIGENT AGENTS: JACK Intelligent Agents or JACK is a framework in Java for multi-agent system development. JACK is a third generation agent platform building on the experiences of the Procedural Reasoning System (PRS) and Distributed Multi-Agent Reasoning System (dMARS) which was built by Intelligent Agents by Agent Oriented Software Pty. Ltd. (AOS). The aim of company is to provide a platform for commercial, industrial and research applications [12, 13]. Major design goals for JACK were: to provide developers with a robust, stable, light-weight product; to satisfy a variety of practical application needs; to ease technology transfer from research to industry; to enable further applied research Bellow is the Features of JACK:  Agent Run-time:  JACK Plan Language (JPL):  Belief-Desire-Intention Model:  Capabilities:  JACK Development Environment (JDE):  Graphical Plans: JACK Object Modeller (JACOB):  Platform Independence: [14] JACK agents are very suitable for larger environments and they are not bound to any specific Agent Communications Language (ACL). It also provides a native light-weight communications

International Conference on Reliability, Infocom Technology and Optimization (ICRITO'2010)

infrastructure for situations where high performance is required [15]. It is very useful in both the Development of distributed systems in software engineer and academic researcher. 6. CONCLUSSION: Multi agent system plays a very vital role in the design of software system especially in the distributed one. The ability of various agents is utilized at a single platform that not only reduces the complexity but also contributes towards toward the specialized approach of development. Multi agent system design has been simplified a great deal with the introduction of frameworks through there are certain areas where work need to be done 7. REFERENCE: [1] Erdem Dur “Multi-Agent Systems” [2] Fabio Bellifemine1, Agostino Poggi, and Giovanni Rimassa,“Developing Multi-agent Systems with JADE,” [3] Victor R. Lesser, ”Encyclopedia of Computer Science, 4th edition”, Pages: 1194 – 1196 [4] FIPA. Foundation for Intelligent Physical Agents, http://www.fipa.org [5] Stefan Poslad, Queen Mary,“History of FIPA,” http://www.fipa.org/subgroups/ROFS-SGdocs/History-of-FIPA.htm [6] Hiroki Suguri, “A Standardization Effort for Agent Technologies The Foundation for Intelligent Physical Agents and Its Activities,” Proceedings of the 32nd Hawaii International Conference on System Sciences - 1999 [7] Fabio Bellifemine, Federico Bergenti, Giovanni Caire and Agostino Poggi, “Jade — A Java Agent Development Framework,” Multi-Agent Programming ,Multiagent Systems, Artificial Societies, and Simulated Organizations, 2005, Volume 15, II, 125-147, DOI: [8]"JADE" http://jade.tilab.com/ [9] Fabio Bellifemine, Giovanni Caire, Agostino Poggi, and Giovanni Rimassa, "JADE: A software framework for developing multi-agent applications. Lessons learned," [10] “Java Agent Development Framework” http://en.wikipedia.org/wiki/Java_Agent_Development_Framework [11] Fabio Bellifemine, Agostino Poggi, Giovanni Rimassa, P. Turci, ”An Object-Oriented Framework to realize Agent System,” [12] Nick Howden, Ralph Rönnquist, Andrew Hodgson, Andrew Lucas, “JACK Intelligent AgentsTM – Summary of an Agent Infrastructure” [13]Paolo Busetta, Ralph Rönnquist, Andrew Hodgson, Andrew Lucas, “Light-Weight Intelligent Software Agents in Simulation” [14] “JACK Intelligent Agents “ http://en.wikipedia.org/wiki/JACK_Intelligent_Agents [15] Paolo Busetta, Ralph Rönnquist, Andrew Hodgson, and Andrew Lucas “JACK Intelligent Agents -Components for Intelligent Agents in Java” Agent Oriented Software Pty. Ltd., Melbourne, Australia

Suggest Documents