intelligent agents review - AMiner

5 downloads 5612 Views 61KB Size Report
Research in the field of software agents has been an active research area in the .... iBundler, which is an agent-based service offered to buying agents to help ...
INTELLIGENT AGENTS REVIEW Khalid Mansour Dept. of Computer Science Zarka Private University, Jordan [email protected]

Ezz Hattab Faculty of Information Technology Al-Ahliyya Amman University, Jordan [email protected]

Emad Abuelrub Dept. of Computer Science Zarka Private University, Jordan [email protected]

Abstract Software intelligent agents, or simply agents, are one of the most important research areas in computer science and information technology. Plenty of work has been done on different aspects of agents development. This paper presents a survey of research activities in the development of intelligent agents. The survey categorizes the agent related work into three main theoretical based areas: (1) Interaction protocols, (2) Design and architecture, (3) Agent applications. The purpose of this categorization is to understand the main research driven directions in this promising field, and help in choosing a particular direction for more in depth studies.

1. Introduction Research in the field of software agents has been an active research area in the last few years, especially after the wide spread of the Internet [1..40]. This emergent field of research has plenty of promising results that may introduce a new revolution on the Internet. Agents and multi-agent systems are being used in many e-commerce applications [3,7,16,23,25,29,32,33]. Searching over the Internet is one of the most frequent activities that are done by Internet users. It is obvious that a lot of work has been done to make searching easy and efficient. However, the main shortcoming is that the task of searching has become a more difficult task, because the web is over flooded with information [20,21,22,28,39]. Intelligent agents are the promising tools that can alleviate this and other obstacles of the Internet [20,21,22,28,39]. This paper tries to filter and categorize trends of research in the field of agents on the Internet and presents some achievements. Three areas of research trends are proposed: Interaction protocols, design and architecture, and applications. This paper is organized as follows: Section 2 addresses the subject of interaction protocols. Section 3 concentrates on the design and architecture of agents. The applications of agents, e.g., ecommerce and web searching are presented in section 4, while section 5 presents the conclusion and proposes future work.

2. Interaction Protocols Software agents need to interact with each other to perform different tasks. Interaction protocols are needed as a means of communication between different agents [6,8,9,17,19,24,30,40]. [6]

745

Focuses on protocols used in Agentis, while [8] formulated his work on the dynamics of social interaction in a multi-agent system within the belief-desire-intention (BDI) paradigm. Task and resource allocation in a computational economy (TRACE) is an adaptive organizational policy for multi-agent systems [9]. TRACE uses two main components: The task allocation protocol (TAP) and the resource allocation protocol (RAP). Adaptive negotiation model is proposed by [17], while a hybrid negotiation mechanism is presented by [19]. Conceptual graph (CG) of [36] is used to describe an interaction protocol [24], while colored petri nets (CPNs) [5,14] is used to facilitate the modeling of concurrent conversations in an integrated style [30]. [35] Proposed a model for designing negotiation agent with two important properties: The first is a market driven strategy, the second is a set of relaxed decision rules. An integrative negotiation mechanism [40], gives agents the option of choosing the mode of negotiation with other agents. [40] Considered the dual concern model [18], and the motivational quantities (MQ) framework [37] to build mapping functions. Table1 summarizes interaction protocols mentioned in this section. Criteria

Approach Agentis framework

Belief-Desire-Intention (BDI)

TRACE Adaptive negotiation model A hybrid negotiation mechanism Interaction Protocols

Conceptual Graph (CG)

Description The framework is used for building interactive multi-agent applications A formal analysis of the dynamics of social interaction in a multi-agent system based on concepts such as roles, commitments and obligations TRACE (Task and Resource Allocation in a Computational Economy) is an adaptive organizational policy for multi-agent systems. A new model for negotiation between agents, supported by the AGM belief revision model The hybrid negotiation method includes both cooperative and competitive negotiations Conceptual graphs are used to describe an interaction protocol that can be used in multiagent systems in a declarative way

The formal properties of Petri nets are used to facilitate the modeling of concurrent conversations in an integrated style. Three mathematical functions are used in Fuzzy Rules addition to a set of fuzzy rules in determining the amount of concession for each negotiation round This mechanism gives agents the option of An integrative negotiation choosing the mode of negotiation with other mechanism agents Table1. A summary of interaction protocols in agent systems

Colored Petri Nets (CPNs)

Reference [6]

[8]

[9] [17]

[19]

[24]

[30]

[35]

[40]

3. Design and Architecture of Agents Design and architecture is a major part in the life cycle of any software development. Many researches focus on this part in agent development [1,4,10,11,13,15,21,26,27,31,35,38]. CORBAbased architecture for lightweight agents (CARLA) is an architecture for Java-based mobile agents to be used on small devices [1]. [4] Proposed an approach for multi-agent systems oriented to robotics applications. This approach is called process for agents specification and im746

plementation (PASSI). ShadowBoard and ShadowFace [10], is a methodology and architecture for rapidly building custom multi-agent applications for a range of agent oriented personal devices. [11] Used user interface agents (UIA) managing information. On demand agent assembling vision is realized by building eHermes, which is a prototype system [13]. [15] Presented a platform for prototyping. Faded information field (FIF) [21], and Message passing interface agent based (MPIAB) [31] are two different architectures used for building agents. AGORA [27], is used in the formation of virtual enterprises. [26] Designed two types of agents to face possible problems in communications. [38] Proposed an architecture for autonomic computing systems. Table2 summarizes designs and architectures mentioned in this section. Criteria

Approach CARLA

PASSI

ShadowBoard & ShadowFace.

Design and architecture of agents

UIA eHermes, an on demand agent assembling A platform for prototyping

Description CARLA (CORBA-based Architecture for Lightweight Agents) project is an architecture for Java-based mobile agents to be used on small devices. PASSI (Process for Agents Specification and Implementation) is used as a design methodology, which is a step-by-step requirement to code methodology for developing multi-agent software The methodology and architecture of ShadowBoard and ShdowFace provides an environment for rapidly building custom multiagent applications for a range of agentoriented personal devices. User interface agents (UIA). They use machine-learning techniques to help a user deal with a particular application eHermes is the prototype system, which is developed as the proof of concept for the “ondemand agent assembling” vision. A platform to be used in prototyping of multiagent systems made of a heterogeneous set of cooperating robots

Faded Information Field (FIF) is a demandoriented architecture and is supported by Push/Pull mobile agents. Design two types of The doubler agent and stand-in agent where agents used for continuing work in case of communication problems AGORA AGORA multi-agent system is used to support the formation of Virtual Enterprise (VE). MPIAB (Message Passing Interface-Agent MPIAB Based) is an agent-based architecture for parallel processing. Autonomic computing Design and build computing systems that systems function like the human nervous system Table2. A summary of the design and architecture of agents FIF

Reference [1]

[4]

[10]

[11] [13]

[15]

[21] [26]

[27]

[31] [38]

4. Applications of Agents Many different applications of software agents or multi-agent systems are recently appeared over the Internet [3,7,20,22,23,25,28,32,34,39]. Kasbah [3] is an agent market place for buying and selling goods. [7] Presented an intelligent agent market place. SportsFinder [20] understands the

747

context of knowledge on the Web pages. SIMPLE [23] is a multi-agent system used in related auctions. SARDINE (system for airline reservations demonstrating the integration of negotiation and evaluation) is an agent-facilitated airline ticket bidding system [25]. [32] Presented the iBundler, which is an agent-based service offered to buying agents to help them determine the optimal choices of received offers based on their constraints and preferences. Apt Decision Agent [34] helps users in finding an apartment that match their need. [39] Introduced the ontology-driven adaptive web information extraction system (Odaies) to help in extracting information from the HTML documents over the Internet. Table3 summarizes agents applications mentioned in this section.

4. Conclusions This paper focuses on the work done in the literature of agents over the Internet. Our work concentrate on three areas: interaction protocols, design and architecture of agents, and applications of agents. The future work may focus on developing and testing a new methodology for interagent interactions. Application Kasbah Intelligent agent market place

SportsFinder

Approach Build a market place. Market place with a market place manager implemented as one of the agents in the market place, it uses dependence theory for the mediation process. An agent without any formal knowledge representation, it understands the context of knowledge on the Web pages.

Information broker system

Developed within the Open Agent Architecture.

SIMPLE

Multi-agent system.

SARDINE

An agent-facilitated bidding system.

Neural based multi-agent system

Based on EĮNet architecture.

iBundler

Multi-agent system.

Apt decision agent OntologyDriven Adaptive Web Information Extraction System (Odaies)

Interactive learning techniques.

Ontology and three algorithms for extracting were used.

Description It is a Web-based agent market place where selling and buying agents meet. The idea of the market place manager is introduced as a facilitator between agents.

Reference [3]

An agent that extracts sporting match results from a wide variety of Web sites. The system provides coordinated access to a heterogeneous collection of structured and semi-structured information. An agent that trade in related auctions. It is an agent system to help in airline ticket bidding, and it presents interfaces for individualized negotiation. The system used to extracts information from repositories containing HTML documents. The system is useful in situations where HTML documents are stored in private networked document repositories, which are not indexed in by search engines. An agent-based service offered to buying agents to help them determine the optimal choices. An agent that learns user preferences in the domain of rental real estate. Extracts information from the HTML documents over the Web.

[20]

Table3. A summary the applications of agents

748

[7]

[22]

[23]

[25]

[28]

[32] [34]

[39]

References [1] ALEKSY, M., Korthaus, A., and Schader, M., CARLA – A CORBA-based Architecture for Lightweight Agents, in: Proceedings of the IEEE/WIC International Conference on Intelligent Agent Technology (IAT’03), Halifax, Canada 2003. [2] AU, R., Yao, M., and Looi, M., Privilege Negotiation Agents for Distributed Authorisation on World Wide Web, in: Proceedings of the IEEE/WIC International Conference on Intelligent Agent Technology (IAT’03), Halifax, Canada 2003. [3] Chavez, A., Maes P., Kasbah: An Agent Marketplace for Buying and Selling Goods, in: Proceedings of the 1st International Conference on the Practical Application of Intelligent Agents and Multi-Agent Technology, London 1996. [4] COSSENTINO, M., Sabatucci L., and Chella A., A Possible Approach to the Development of Robotic MultiAgent Systems, in: Proceedings of the IEEE/WIC International Conference on Intelligent Agent Technology (IAT’03), Halifax, Canada 2003. [5] COST, S., Chen, Y., Finin, T., Labrou, Y., and Peng, Y., Using Colored Petri Nets for Conversation Modeling, Issues in Agent Communication, Lecture Notes in AI, Springer-Verlag, Berlin 2000. [6] D'INVERNO, M., Kinny D., and Luck M., Interaction Protocols in Agentis, in: ICMAS'98, Third International Conferenceon Multi-Agent Systems, IEEE Computer Society, Paris, France 1998. [7] ESMAHI, L., and Dini P., Toward an Open Virtual Market Place for Mobile Agents, in: IEEE 8th International Workshops on Enabling Technologies: Infrastructure for Collaborative Enterprises, Palo Alto, California 1999. [8] FASLI, M., Social Interactions in Multi-agent Systems: A Formal Approach, in: Proceedings of the IEEE/WIC International Conference on Intelligent Agent Technology (IAT’03), Halifax, Canada 2003. [9] FATIMA, S. S., and Wooldridge M., Adaptive Task and Resource Allocation in Multi-agent Systems, in: Proceedings of the Fifth International Conference on Autonomous Agents, 2001. [10] GOSCHNICK, S., and Sterling L., An Agent-based Digital Self in a 24x7 Web Services World: Architecture and Implementation, in: Proceedings of the IEEE/WIC International Conference on Intelligent Agent Technology (IAT’03), Halifax, Canada 2003. [11] HELMY, T., Amamiya S., Mine T., and Amamiya M., A New Approach of the Collaborative User Interface Agents, in: Proceedings of the IEEE/WIC International Conference on Intelligent Agent Technology (IAT’03), Halifax, Canada 2003. [12] HERMANS, B., Intelligent Software Agents on the Internet: An Inventory of Currently Offered Functionality in the Information Society & a Prediction of (Near-) Future Development, MS Thesis, Tilburg University, 1996. [13] JAYAPUTERA, G., Loke S., and Zaslavsky A., Mission Impossible? Automatically Assembling Agents from High-Level Task Descriptions, in: Proceedings of the IEEE/WIC International Conference on Intelligent Agent Technology (IAT’03), Halifax, Canada, 2003. [14] JENSEN, K., Coloured Petri Nets – Basic Concepts, Analysis Methods and Practical Use, Springer-Verlag, Berlin 1992. [15] JENSEN, L., Kristensen B., and Demazeau Y., FLIP: A Platform to Integrate Embodied Agent Technology, in: Proceedings of the IEEE/WIC International Conference on Intelligent Agent Technology (IAT’03), Halifax, Canada 2003. [16] LAU, H., and Lim W., Multi-agent Coalition Via Autonomous Price Negotiation in a Real-Time Web Environment, in: Proceedings of the IEEE/WIC International Conference on Intelligent Agent Technology (IAT’03), Halifax, Canada 2003. [17] LAU, R., Essam B., Chan S., and Huang Z., Belief Revision for Adaptive Negotiation Agents, in: Proceedings of the IEEE/WIC International Conference on Intelligent Agent Technology (IAT’03), Halifax, Canada 2003. [18] LEWICKI, R. J., and Litterer J. A., Negotiation. Richard D. Irwin Inc, Homewood, Illinois 1985. [19] LI, X., and Soh L., The Use of Hybrid Negotiation in Resource Coordination Among Agents, in: Proceedings of the IEEE/WIC International Conference on Intelligent Agent Technology (IAT’03), Halifax, Canada 2003. [20] LISTER, K., and Sterling L., Tasks as Context for Intelligent Agents, in: Proceedings of the IEEE/WIC International Conference on Intelligent Agent Technology (IAT’03), Halifax, Canada 2003. [21] LU, X., and Mori K., Autonomous Information Services Integration and Allocation in Agent-Based Information Service System, in: Proceedings of the IEEE/WIC International Conference on Intelligent Agent Technology (IAT’03), Halifax, Canada 2003. [22] MARTIN, D., Oohama H., Moran D., and Cheyer A., Information Brokering in an Agent Architecture, in: Proceedings of the Practical Application of Intelligent Agents and Multi-agent Technology (PAAM 97), 1997. [23] MILIDIU, R., Melcop T., Liporace F., and de Lucena C., SIMPLE - A Multi-Agent System for Simultaneous and Related Auctions, in: Proceedings of the IEEE/WIC International Conference on Intelligent Agent Technology (IAT’03), Halifax, Canada 2003.

749

[24] MINEAU, G., Representing and Enforcing Interaction Protocols in Multi-Agent Systems: an Approach Based on Conceptual Graphs, in: Proceedings of the IEEE/WIC International Conference on Intelligent Agent Technology (IAT’03), Halifax, Canada 2003. [25] MORRIS, J. and Maes P., Sardine: An Agent-facilitated Airline Ticket Bidding System, in: Fourth International Conference on Autonomous Agents (Agents 2000), Barcelona, Catalonia, Spain 2000. [26] PECHOUCEK, M., Dobisek M., Lazansky J., and Marlk V., Inaccessibility in Multi-Agent Systems, in: Proceedings of the IEEE/WIC International Conference on Intelligent Agent Technology (IAT’03), Halifax, Canada 2003. [27] PETERSEN, S., Rao J., and Matskin M., Virtual Enterprise Formation with Agents – an Approach to Implementation, in: Proceedings of the IEEE/WIC International Conference on Intelligent Agent Technology (IAT’03), Halifax, Canada 2003. [28] PILATO, G., Vitabile S., Vassallo G., Conti V., and Sorbello F., A Neural Multi-Agent Based System for Smart Html Pages Retrieval, in: Proceedings of the IEEE/WIC International Conference on Intelligent Agent Technology (IAT’03), Halifax, Canada 2003. [29] PURVIS, M., Cranefield, S., Nowostawski, M., and Carter, D., Opal: A Multi-Level Infrastructure for AgentOriented Software Development, Information Science Discussion Paper Series, No. 2002/01, ISSN 1172-6024, University of Otago, Dunedin, New Zealand 2002. [30] PURVIS, M., Nowostawski M., Oliveira M., and Cranefield S., Multi-Agent Interaction Protocols for eBusiness, in: Proceedings of the IEEE/WIC International Conference on Intelligent Agent Technology (IAT’03), Halifax, Canada 2003. [31] RAHIMI, S., Narayanan A., and Sabharwal M., MPIAB: A Novel Agent Architecture for Parallel Processing, in: Proceedings of the IEEE/WIC International Conference on Intelligent Agent Technology (IAT’03), Halifax, Canada 2003. [32] RODRIGUEZ-AGUILAR, J., Giovanucci A., Reyes-Moro A., Noria F., and Cerquides J., Agent-based Decision Support for Actual-world Procurement Scenarios, in: Proceedings of the IEEE/WIC International Conference on Intelligent Agent Technology (IAT’03), Halifax 2003. [33] ROTHKOPT, M. H., Pekec A., and Harstad R. M., Computationally Manageable Combinatorial Auctions, Management Science, 8(44): 1131-11147, 1995. [34] SHEARIN, S., and Lieberman H., Intelligent Profiling by Example, in: Proceedings of the International Conference on Intelligent User Interfaces (IUI 2001), Sante Fe, New Mexico 2001. [35] SIM, K., and Wang S., Designing Flexible Negotiation Agent with Relaxed Decision Rules, in: Proceedings of the IEEE/WIC International Conference on Intelligent Agent Technology (IAT’03), Halifax, Canada 2003. [36] SOWA, J. F., Conceptual Structures: Information Processing in Mind and Machine, Addison-Wesley, 1984. [37] WAGNER, T., and Lesser V., Evolving Real-time Local Agent Control for Large-scale Mass, in: J. Meyer and M. Tambe, (editors), Intelligent Agents VIII (Proceedings of ATAL01), Lecture Notes in Artificial Intelligence, Springer-Verlag, Berlin 2002. [38] TIANFIELD, H., Multi-Agent Autonomic Architecture and Its Application in E- Medicine, in: Proceedings of the IEEE/WIC International Conference on Intelligent Agent Technology (IAT’03), Halifax, Canada 2003. [39] ZHANG, N.,Chen H., Wang Y., Cheng S., and Xiong M., Odaies: Ontology-driven Adaptive Web Information Extraction System, in: Proceedings of the IEEE/WIC International Conference on Intelligent Agent Technology (IAT’03), Halifax, Canada 2003. [40] ZHANG, X., Lesser V., and Wagner T., Integrative Negotiation In Complex Organizational Agent Systems, in: Proceedings of the IEEE/WIC International Conference on Intelligent Agent Technology (IAT’03), Halifax, Canada 2003.

750