Anne Håkansson, Ronald Hartung, and Ngoc Thanh Nguyen (Eds.) Agent and Multi-agent Technology for Internet and Enterprise Systems

Transcription

1 Anne Håkansson, Ronald Hartung, and Ngoc Thanh Nguyen (Eds.) Agent and Multi-agent Technology for Internet and Enterprise Systems

2 Studies in Computational Intelligence, Volume 289 Editor-in-Chief Prof. Janusz Kacprzyk Systems Research Institute Polish Academy of Sciences ul. Newelska Warsaw Poland kacprzyk@ibspan.waw.pl Further volumes of this series can be found on our homepage: springer.com Vol Johann M.Ph. Schumann and Yan Liu (Eds.) Applications of Neural Networks in High Assurance Systems, 2009 ISBN Vol Francisco Fernández de de Vega and Erick Cantú-Paz (Eds.) Parallel and Distributed Computational Intelligence, 2009 ISBN Vol Zong Woo Geem Recent Advances In Harmony Search Algorithm, 2009 ISBN Vol Janusz Kacprzyk, Frederick E. Petry, and Adnan Yazici (Eds.) Uncertainty Approaches for Spatial Data Modeling and Processing, 2009 ISBN Vol Carlos A. Coello Coello, Clarisse Dhaenens, and Laetitia Jourdan (Eds.) Advances in Multi-Objective Nature Inspired Computing, 2009 ISBN Vol Fatos Xhafa, Santi Caballé, Ajith Abraham, Thanasis Daradoumis, and Angel Alejandro Juan Perez (Eds.) Computational Intelligence for Technology Enhanced Learning, 2010 ISBN Vol Zbigniew W. Raś and Alicja Wieczorkowska (Eds.) Advances in Music Information Retrieval, 2010 ISBN Vol Dilip Kumar Pratihar and Lakhmi C. Jain (Eds.) Intelligent Autonomous Systems, 2010 ISBN Vol Jacek Mańdziuk Knowledge-Free and Learning-Based Methods in Intelligent Game Playing, 2010 ISBN Vol Filippo Spagnolo and Benedetto Di Paola (Eds.) European and Chinese Cognitive Styles and their Impact on Teaching Mathematics, 2010 ISBN Vol Radomir S. Stankovic and Jaakko Astola From Boolean Logic to Switching Circuits and Automata, 2010 ISBN Vol Manolis Wallace, Ioannis E. Anagnostopoulos, Phivos Mylonas, and Maria Bielikova (Eds.) Semantics in Adaptive and Personalized Services, 2010 ISBN Vol Chang Wen Chen, Zhu Li, and Shiguo Lian (Eds.) Intelligent Multimedia Communication: Techniques and Applications, 2010 ISBN Vol Robert Babuska and Frans C.A. Groen (Eds.) Interactive Collaborative Information Systems, 2010 ISBN Vol Husrev Taha Sencar, Sergio Velastin, Nikolaos Nikolaidis, and Shiguo Lian (Eds.) Intelligent Multimedia Analysis for Security Applications, 2010 ISBN Vol Ngoc Thanh Nguyen, Radoslaw Katarzyniak, and Shyi-Ming Chen (Eds.) Advances in Intelligent Information and Database Systems, 2010 ISBN Vol Juan R. González, David Alejandro Pelta, Carlos Cruz, Germán Terrazas, and Natalio Krasnogor (Eds.) Nature Inspired Cooperative Strategies for Optimization (NICSO 2010), 2010 ISBN Vol Roberto Cipolla, Sebastiano Battiato, and Giovanni Maria Farinella (Eds.) Computer Vision, 2010 ISBN Vol Alexander Bolshoy, Zeev (Vladimir) Volkovich, Valery Kirzhner, and Zeev Barzily Genome Clustering, 2010 ISBN Vol Dan Schonfeld, Caifeng Shan, Dacheng Tao, and Liang Wang (Eds.) Video Search and Mining, 2010 ISBN Vol I-Hsien Ting, Hui-Ju Wu, Tien-Hwa Ho (Eds.) Mining and Analyzing Social Networks, 2010 ISBN Vol. 289.AnneHåkansson, Ronald Hartung, and Ngoc Thanh Nguyen (Eds.) Agent and Multi-agent Technology for Internet and Enterprise Systems, 2010 ISBN

3 Anne Håkansson, Ronald Hartung, and Ngoc Thanh Nguyen (Eds.) Agent and Multi-agent Technology for Internet and Enterprise Systems 123

4 Prof. Anne Håkansson Department of Computer and Systems Sciences Forum 100, SE Kista Sweden Dr. Ronald Hartung Division of Computer and Information Sciences Franklin University Columbus Ohio USA Prof. Ngoc Thanh Nguyen Institute of Computer Science Wroclaw University of Technology Str. Wyb. Wypsianskiego Wroclaw Poland ISBN e-isbn DOI / Studies in Computational Intelligence ISSN X Library of Congress Control Number: c 2010 Springer-Verlag Berlin Heidelberg This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilm or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permission for use must always be obtained from Springer. Violations are liable to prosecution under the German Copyright Law. The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. Typeset & Cover Design: Scientific Publishing Services Pvt. Ltd., Chennai, India. Printed on acid-free paper springer.com

5 Preface Research in multi-agent systems offers a promising technology for problems with networks, online trading and negotiations but also social structures and communication. This is a book on agent and multi-agent technology for internet and enterprise systems. It represents the extension of agent systems into an applications domain for business systems. The book is a pioneer in the combination of the fields and is based on the concept of developing a platform to share ideas and presents research in technology in the field and application to real problems. The chapters range over both applications, illustrating the possible uses of agents in an enterprise domain, and design and analytic methods, needed to provide the solid foundation required for practical systems. The organization of the chapters is arranged into several areas, however the ideas in the chapters overlap between the analytic work and the applications themselves. The first three chapters of this book present samples of efforts of providing agents and multi-agent technology in distributed environments. It discusses handling agents in networks, and architectures, as well as, service fusion in these environments. Based on some technology in distributed systems, applications for are provided for domain knowledge, trading, information handling and network analysis. From these chapters, the focus moves towards analytical methods for organizing and managing large agent systems. These conduct logical analyses, multi-agent learning and automatically creating hierarchies. The following chapters of the book present formal analyses using modal languages and technical solutions to problems with large state spaces with a large set of information. These chapters also present solutions to handle cognitive architectures, mobile agents, personalized ontologies, e-commerce, multimedia and interfaces. The contributors to the book were carefully selected authors with peer-reviewed chapters. The authors are academics, educators, and practitioners who are doing research in agents and multi-agents technology, as well as, internet and enterprise systems, to meet, improve, and overcome challenges of the internet and enterprise realm. The agent and multi-agent technology is applied in implementing real-life, exciting developments associated with the need to eliminate problems of the organizations. The opening chapter of the book, Chapter 1, is titled Performance models for multi-agent systems and mobile agent network and authored by Mario Kusek, Kresimir Jurasovic, Ignac Lovrek, Vjekoslav Sinkovic, and Gordan Jezic. The authors present performance models for distributed multi-agent systems and mobile agent networks applied to the Internet and enterprise systems. The concept of a mobile agent network extends the notion of a multi-agent system to a networked and distributed environment, supporting agent hosting, execution, communication and migration, as well as enabling interaction among service users and service providers. In this chapter, the authors elaborate the problem of establishing and configuring agent-based distributed systems, including agent service capability, allocation of

6 VI Preface services to agents, and assignment of agents to system nodes. The authors also discuss strategies for distributing services to agents and assigning agents to nodes, and optimization of the distribution. The following chapter, Chapter 2, titled Service Oriented Architecture and Agents: parallels and opportunities is authored by Ronald L Hartung. The chapter presents a synergy between Service Oriented Architecture (SOA) and multi-agent systems by comparing them and emphasizing their support for each other. In the chapter, the author introduces the reader to several theses surrounding SOA and agent technical approaches, as well as solutions using agent techniques and agent systems to support SOA. The author gives a review of the principles that underlie SOA and looks at some agent system implementations, from which, analysis of possible architectural styles for agent systems is explored. Architectural styles are similar to patterns and are a useful tool for enterprise architects and, therefore, these are easily extended into agent architectures. From some of the work in agent systems, extensions into the SOA architectures are elaborated. Conclusions of the chapter show that SOA can provide a strong framework for agent systems development and deployment. The next chapter, Chapter 3, Service Fusion in Mobile Contexts, is authored by Jun Miao Shen, Michael O'Grady, and Gregory O'Hare In this chapter, the authors use the agent paradigm to extend services beyond the Internet on to mobile devices. The authors reflect on some of the pertinent issues necessary for the agents constructing a viable platform for service delivery to mobile users, including Agent Oriented Software Engineering (AOSE). Developments in AOSE and agent frameworks result in an increasing range of options for those considering harnessing the characteristics of agents for their applications and services. The discussion is illuminated by a systematic description of EasyLife a framework that harnesses the agent paradigm to construct and deliver services to mobile users. EasyLife demonstrates the viability of the agent paradigm for integrating heterogonous information sources for service provision to mobile users. The robustness of the agent platform is verified through the use of two separate agent toolkits being harnessed to deliver EasyLife services. The chapter that follows, Chapter 4, is titled Environmental Impact Assessment by Multi-Agent Systems and is authored by Marina V. Sokolova and Antonio Fernández- Caballero. The authors propose a generalized multi-agent system technology (MAST) that facilitates the workflow of multi-agent system s planning and implementation in decision support systems (DSS) applied to Environmental Impact Assessment (EIA). This research has already showed promising results in the field of Environmental Impact Assessment and implementation of decision support systems, and the authors demonstrate an extension of a multi-agent decision support system to new problem domains by changing the domain ontology. The authors introduce and discuss how the functional organization and the roles, executed in the MAS, conform to the logical sequence of the data transformation flow, which includes information retrieval and fusion, data preprocessing, data mining and modeling, and simulation and decision making. The next chapter, Chapter 5, is titled Multiagent Forex Trading System and authored by Rui Pedro Barbosa and Orlando Belo.

7 Preface VII The authors of this chapter use multi-agent systems to implement currency automated trading systems, which is a specific case of e-commerce. The work analyzes the intelligent agent s ability to act autonomously as well as adapt and interact with the environment. The authors introduce and describe the implementation of a multi-agent system that can achieve results trading in the Forex market. The implementation includes a set of autonomous currency trading agents, using an infrastructure that consists of an ensemble of classification and regression models, a case-based reasoning system, and an expert system. The authors also demonstrate how investment diversification can be used to reduce the trading risk, and show how enabling the agents to communicate among themselves can help reduce the trading costs and the capital requirements. The system can lessen the impact of the trading costs to increase the total profit by integrating the agents in a multi-agent system in which the agents communicate with each other before opening new trades. The Chapter that follows, Chapter 6, is titled Information coordination using Meta- Agents in Information Logistics Processes and authored by Mats Apelkrans and Anne Håkansson. The chapter presents a work that shows the combination of multi-agent system and information logistic process (ILP) to facilitate the information workflow in business processes. The authors combine theories from multi-agent systems and Information Logistics (IL) to coordinate and deliver information at the right time and to the right place at an acceptable cost. The multi-agent system contains agents and meta-agents that search, coordinate, and deliver the information. Information Logistics is, on the other hand, an information product that becomes accessible and delivered to the information receiver who can make use of the information. The agents and the metaagents support the information flow between the manufacturing company, its customers, and its suppliers in the form of orders, requirements, and production plans, which can speed up the information exchange and optimize the production cost. The next chapter, Chapter 7, is titled Social Network Analysis Based Approach to Trust Modeling for Autonomous Multi-agent Systems and authored by Grzegorz Kolaczek. The chapter extends social network analysis to multi-agent systems in small-world on a scale-free network to generate trust relationships. The author presents a general concept about trust modelling and management using a network based analysis of relations between autonomous agents and an analysis method for autonomous multiagent systems. The method is based on the assumption that multi-agent systems constitute social networks and the network specific parameters can be used to evaluate the trust levels. The author describes how metrics like centrality and clustering coefficient used in the description of these types of networks can be used to model trust relation between agents. Thus, the parameters describing small-world or scale-free networks can be also used to analyze the position and relations among agents within their societies. The following chapter, Chapter 8, titled Logical Analysis of Multi-Agent Systems is authored by Sergey Babenyshev and Vladimir Rybakov. The independent operation of multi-agent systems increases the difficulty of understanding and analysing the agents operations. In this chapter, the authors discuss several interdependent theoretical tools from modern mathematical logic and computer science for formalizing the informal notion of multi-agent systems. The

8 VIII Preface authors explain how the modal propositional language may be used to specify the behaviour of the processes, and formally introduce the Kripke semantics for modal languages. The authors also discuss different techniques of combining simple modal languages into complex ones, more suitable for practical purposes, and conclude with an extended example, illustrating the major points of the above-mentioned techniques. The next chapter, Chapter 9, is titled Multi-agent systems and large state spaces and authored by Yann-Michaël De Hauwere, Peter Vrancx, and Ann Nowé. When a multi-agent system uses reinforcement learning, the size of the state space inhibits effective learning. A solution explored by the authors of this chapter is Generalized Learning Automata (GLA). The GLA can learn generalisations over a large number of states, without needing to store large amounts of samples, perform computational intensive statistical tests, or build a model of the environment. GLA can help take the correct actions in large, unknown environments. The authors introduce a general framework for multi-agent learning, where learning happens on two separate layers, and agents learn when to observe each other. Within this framework, the authors introduce a new algorithm, called 2observe, which uses a GLA approach to distinguish between high-risk states where the agents have to take each other s presence into account and low risk states where they can act independently. The algorithm is applied to a grid-world problem since it has similarities to some real-world problems, such as autonomous robot control. The chapter that follows, Chapter 10, is titled Automatically creating hierarchies of agents and meta-agents using clustering and authored by Anne Håkansson. This chapter describes generating hierarchies of agents to control and master problem solving in multi-agent systems. The hierarchies are established automatically in response to the problem solving. The author presents how to create hierarchies of agents automatically from the result of searching for information. The agents, both intelligent software agents and meta-agents, are represented in a hierarchy to cluster similar agents and optimize locality. The commonality between the agents represents semantics, and constitutes information chains for solutions. The clustering can support avoiding the unsuccessful paths and makes the search efficient. In addition, the hierarchy supports reasoning at different levels in hierarchy by dividing the overall task for the agents in the hierarchy. As a test environment, the author uses the web, since it can produce a vast amount of information that needs structuring in order to be manageable and accessible. The results of the search infrastructure are automatically cast into a more manageable and understandable hierarchy. The following chapter, Chapter 11, titled A Role-based Cognitive Architecture for Multi-Agent Teaming is authored by Jinsong Leng, Chee Peng Lim, Jiuyong Li, Dongguang Li and Lakhmi Jain. The authors in this chapter introduce a development of a role-based architecture to enable teaming and collaboration to organise problem solving in a multi-agent system. Multi-agent teaming can be constructed in a layered architecture with individual agent behaviours and team behaviours where the architecture uses Belief, Desire, and Intension (BDI) as an argumentation framework. The role-based BDI framework facilitates optimization problems at the team level such as competitive, cooperation, and coordination problems. This BDI framework is extended on the commercial agent software development environment known as JACK Teams. The layered architecture has been used to group the agents competitive and cooperative

9 Preface IX behaviours. The chapter includes a presentation of the use of reinforcement learning techniques to learn different behaviours through experience. These issues have been investigated and analysed using a real-time, 2D simulation environment known as Soccer-Bots. The next chapter, Chapter 12, is titled Mobile Agents and Eventually Perfect Predicate Detection: an Intelligent General Approach to Monitoring Distributed Systems and authored by Chunbo Chu and Monica Brockmeyer. In this chapter, the authors introduce a monitoring approach to oversee a multiagent system using a form of prediction detection that is guaranteed to reach correct states. The authors present an application of mobile agents to solve distributed predicate detection in realistic applications. As a solution, the authors make an extension of the study of predicate detection in failure-prone systems. The extension includes applying the mobile agent technique to implement an eventually accurate failure detection sequencer in a Finite Average Response (FAR) model. In this chapter, the authors propose a general-purpose solution for predicate detection in partially synchronous distributed systems that contributes to predicate detection in large-scale distributed systems with crash failures and limited synchrony. The chapter that follows, Chapter 13, is titled Personalized Ontology-based Recommender Systems for Multimedia Objects and authored by Krzysztof Juszczyszyn, Przemyslaw Kazienko and Katarzyna Musial The chapter demonstrates the application of ontology formation to construct personalised recommendations in the multimedia domain. The Authors propose a framework for recommendation of multimedia objects based on processing of individual ontologies. The technique compares ontologies as whole structures to assess similarity between multidimensional profiles of users and multimedia objects in the recommender system. The ontologies provide the comprehensive view of the information gathered in the multimedia sharing system. As a result, the executed recommendation process takes into account similarities calculated both between objects and users ontologies, which reflect the social and semantic features existing in the system. The ontologies, which are close to the current context, provide a list of suggestions presented to the user. Each user in the system possesses its own Personal Agent that performs all necessary online tasks. Personal Agents co-operate each other and enrich lists of possible recommendations. The system is developed for the Flickr multimedia sharing system. The next chapter, Chapter 14, is titled Toward Emotional E-Commerce and authored by Nicolae Jascanu, Veronica Jascanu and Severin Bumbaru. In this chapter, the authors apply emotional models to negotiate in E-commerce applications and introduce the main actors of the multi-agent system. The authors propose a multi-agent system for electronic commerce that integrates emotional models for each one of the three agents: the customer, the supplier, and the community. Here the authors are using a continuous approach as a model for emotions based on the seminal work of James Russell. The novelty of the work in this chapter is the mapping of the emotional model to the electronic commerce field. For this, the consumer and supplier behaviors are translated in terms of emotional models and an algorithm to process negotiation configurations. By including an emotional community agent in the negotiation to support the customer and the supplier, the work lays out a foundation for emotional based argumentation.

10 X Preface The last chapter of this selection, Chapter 15, is titled Web-based Affective Human-agent Interaction Generation and authored by Xia Mao and Zheng Li. In this chapter, the authors describe the construction of perceptive user interfaces using agent technologies. This is a higher level of interface for agent applications especially in the Internet and extends the usability of agents. The authors describe the architecture of Multimodal Interaction Markup Language (MIML), a XML-based markup language for scripting the web-based intelligent interaction of virtual agents. MIML cannot only simply respond with a pre-defined utterance dependent on the user's input, but also produces dynamic responses, which are determined by the user's emotional state. Unlike existing related languages, MIML can script not only the presentations of virtual agents, but also their affective capability. The authors also describe facial expression recognition, and speech emotion recognition. They also describe emotional speech synthesis ActiveX controllers and illustrate one scenario that instantiates the affective web-based human-agent interaction scripted by MIML. With the MIML design, web-based affective interaction can be described and generated easily. By the actual web-based intelligent interaction implemented with MIML, the tutoring system shows how to mark up the interaction system that features a virtual agent with intelligent behaviour. AnneHåkansson Stockholms University, Kista, Sweden Ronald Hartung Franklin University, Columbus, Ohio, USA Ngoc Thanh Nguyen Wroclaw University of Technology, Wroclaw, Poland

11 Table of Contents Performance Models for Multi-agent Systems and Mobile Agent Network... 1 Mario Kusek, Kresimir Jurasovic, Ignac Lovrek, Vjekoslav Sinkovic, and Gordan Jezic Service Oriented Architecture and Agents: Parallels and Opportunities Ronald L. Hartung Service Fusion in Mobile Contexts Jun Miao Shen, Michael J. O Grady, and Gregory M.P. O Hare Environmental Impact Assessment by Multi-Agent Systems Marina V. Sokolova and Antonio Fernández-Caballero Multi-Agent Forex Trading System Rui Pedro Barbosa and Orlando Belo Information Coordination Using Meta-agents in Information Logistics Processes Mats Apelkrans and Anne Håkansson Social Network Analysis Based Approach to Trust Modeling for Autonomous Multi-agent Systems Grzegorz Ko laczek Logical Analysis of Multi-agent Kripke Structures Sergey Babenyshev and Vladimir Rybakov Multi-Agent Systems and Large State Spaces Yann-Michaël De Hauwere, Peter Vrancx, and Ann Nowé Automatically Creating Hierarchies of Agents and Meta-agents Using Clustering Anne Håkansson A Role-Based Cognitive Architecture for Multi-Agent Teaming Jinsong Leng, Chee Peng Lim, Jiuyong Li, Dongguang Li, and Lakhmi Jain Mobile Agents and Eventually Perfect Predicate Detection: An Intelligent General Approach to Monitoring Distributed Systems Chunbo Chu and Monica Brockmeyer

12 XII Table of Contents Personalized Ontology-Based Recommender Systems for Multimedia Objects Krzysztof Juszczyszyn, Przemys law Kazienko, and Katarzyna Musia l Toward Emotional E-Commerce Nicolae Jascanu, Veronica Jascanu, and Severin Bumbaru Web-Based Affective Human-Agent Interaction Generation Xia Mao and Zheng Li Author Index