SIWN Abstracts Index



sai: cosiwn.2008.06.161

Formal Refinement Model of OSGi based System

Jiankun Wu, Linpeng Huang, Dejun Wang

Communications of SIWN, Vol. 4, June 2008, pp. 1-5

Abstract: An ASM-based formal refinement model for OSGi based system is presented in this paper. The model not only provides the refinement methods but also gives the refinement verification between two ASMs. Finally, the final model that specifies the constraint of OSGi is obtained. The goal of the model is the development of correct software for safety critical applications. From requirement to implement, the stepwise refinement will be under rigorous proving. The trust of the objective system will be guaranteed by the refinement model.

Keywords: refinement, Abstract State Machine (ASM), OSGi, verification, coupling invariant.


sai: cosiwn.2008.06.162

FUNGO: An Algorithm to Derive Local Interaction Protocols from WS Choreographies

Roberto Furnari

Communications of SIWN, Vol. 4, June 2008, pp. 6-10

Abstract: This paper presents FUNGO, an algorithm that, given a complex service definition (as a Web Service choreography model), checks if it can be correctly implemented by the local models of the single services participating in it. When the local models do not fulfil some synchronization constraints defined in the global choreography, the algorithm adjusts them by adding synchronization messages that guarantee that their execution actually implements the interactions required by the choreography definition. The algorithm is part of the PERCHE choreography registry, that extends the idea of UDDI registry to the management of web service choreographies.

Keywords: Choreography, Web Services, SOA, Service Interoperability, Business Process Integration, Synchronization.


sai: cosiwn.2008.06.163

BDI Agent-Oriented Design for Distributed Intrusion Detections

Dayong Ye, Quan Bai and Minjie Zhang

Communications of SIWN, Vol. 4, June 2008, pp. 11-17

Abstract: Many research works on application of intelligent agents in intrusion detection have been done in the past decade. However, they mostly focus on designing a general architecture of agent-based intrusion detection system rather than the intelligent behaviour of agents. In this paper, the design and implementation of Belief-Desire-Intention agents in a Mobile Agent-based Peer-to-Peer Distributed Intrusion Detection Framework (MADIDF) are introduced. In addition, an example which demonstrates the interaction among different agents on different hosts when there is a suspicious distributed attack is also presented. Compared to current related research, MADIDF implements a novel P2P distributed intrusion detection framework, so it can detect attack not only on a single host but also in a distributed domain.

Keywords: BDI Agent, Distributed Intrusion Detection.


sai: cosiwn.2008.06.164

Semantics Supported Access Authorization Based on Decentralized Architecture

Mohammad M. R. Chowdhury, Josef Noll and Najeeb Elahi

Communications of SIWN, Vol. 4, June 2008, pp. 18-24

Abstract: Access authorization is one of the ways to provide information security and privacy assurance. This paper proposes an access authorization mechanism in organizations based on decentralized architecture. It includes decentralized organizational structures containing various attributes which are formally represented using the Web Ontology Language. Access authorization decisions are derived through semantic rules and queries. The proposed architecture is compared with its centralized counterpart and with the relational database approach from the computational complexity, management and maintenance point of views.

Keywords: access authorization, decentralized architecture, ontology, rule, semantic.


sai: cosiwn.2008.06.165

Research of Modeling and Application on Multi-Agent System Using Petri Nets

Liang Chen, Zhijun Yan and Jiajia Liu

Communications of SIWN, Vol. 4, June 2008, pp. 25-28

Abstract: In a multi-agent system, how agents accomplish a goal task is usually specified by multi-agent plans built from basic actions (e.g. operators) of which the agents are capable. The processes of modeling multi-agent system are hindered by the agent character like autonomy, social ability, reactivity, pre-activeness etc. To make up the flaw of former modeling methods, this paper presents a formal approach called AOPN (Agent Oriented Petri Net) for analyzing and modeling multi-agent behaviors using Petri nets. The formalized description of AOPN model is given and an example is presented to make the model be understood more clearly.

Keywords: multi-agent system, Petri nets, modeling


sai: cosiwn.2008.06.166

Colored Petri Net Based Scheduling for Agent-Based Grid Systems

Quan Bai and Minjie Zhang

Communications of SIWN, Vol. 4, June 2008, pp. 29-33

Abstract: Scheduling is one of the most important issues in the area of Grid computing. How to efficiently schedule computing jobs to the distributed, heterogeneous and dynamic computing resources is a key challenge in most Grid implementations. As a powerful artificial intelligence technology, intelligent agent shows considerable promise as a new paradigm for building open systems. In this paper, we target at agent based Grid Computing Systems, and present a Colored Petri Net based strategy for Grid scheduling. Comparing to the existing scheduling strategies, the new strategy is leveraged by the intelligent agents and Colored Petri Nets, therefore is more efficient and flexible for open environments.

Keywords: Grid, multi-agent system, colored Petri net, scheduling, resource management.


sai: cosiwn.2008.06.167

Discovering and Exchanging Information About Users in a SOA Environment

Federica Cena and Roberto Furnari

Communications of SIWN, Vol. 4, June 2008, pp. 34-38

Abstract: This work describes a framework providing a complete solution for the discovery and exchange of information about users in an open dynamic environment. The framework is built on a SOA-based environment where standardWeb Service technologies are exploited to achieve syntactic interoperability and Semantic Web languages are used to achieve semantic interoperability. A central enhanced-UDDI registry is used to support the discovery and cooperation between applications. The exchange of information between systems takes place in a P2P way, by means of atomic communication (if systems share the same knowledge model) or by means of negotiation techniques based on dialogue (in case of different knowledge models).

Keywords: User Model Interoperability, Service-Oriented Architecture, Web Services, Semantic Web, Dialogue.


sai: cosiwn.2008.06.168

A Condition Formula Search

Toshio Kodama, Tosiyasu L. Kunii and Yoichi Seki

Communications of SIWN, Vol. 4, June 2008, pp. 39-44

Abstract: Cyberworlds are distributed systems where data and their dependencies are constantly changing and evolving. In such business application systems, combinatorial explosion happens because you must modify schemas and application programs whenever schemas change, if you use existing techniques. To solve the problem, we have developed a data processing system called Cellular Data System (CDS) based on the cellular model, which is considered the most appropriate to model cyberworlds, using an algebra system called Formula Expression. In this paper, we design and implement a condition formula and its processing maps as an important function in CDS. A condition formula search is a very effective measure when you want to analyze data in cyberworlds without losing consistency in the entire system, since you can search for the data you want without changing application programs, if you employ a condition formula search. That is, a condition formula search is an analysis measure for the worlds under the assumption of frequent changes of schemas. Therefore, if you use CDS, the development process is completely different from the general one, since we do not have to design business specification clearly at requirements analysis. In addition, we demonstrate the effectiveness of a condition formula search by taking up an example of a photo file management system.

Keywords: cyberworlds, cellular model, formula expression, topological space, cellular space, condition formula


sai: cosiwn.2008.06.169

Fuzzy-Controlled Intelligent and Robust Multicast (FIRM) Routing in Mobile Ad Hoc Networks

Anuradha Banerjee and Paramartha Dutta

Communications of SIWN, Vol. 4, June 2008, pp. 45-50

Abstract: In this article we present a protocol for fuzzycontrolled, intelligent and robust multicast (FIRM) routing in mobile ad hoc networks, which establishes and maintains a shared mesh structure for each multicast group, without requiring pre-assignment of core to groups. Using simulations in Qualnet 3.5, we compare FIRM with ODMRP and MAODV which are state-of-the-art representatives of mesh-based and tree-based multicast routing protocols respectively, for mobile ad hoc networks. The results from a wide range of scenario show that FIRM attains higher packet delivery ratio at much lesser cost than ODMRP and MAODV.

Keywords: Ad Hoc networks, routing, multicasting, core, mesh, tree.


sai: cosiwn.2008.06.170

Design and Realization of Linux Based Wireless Data Acquisition System

Y. Cheng, Y. Fun Hu, P. Jiang, S. Brown and A. Metcalfe

Communications of SIWN, Vol. 4, June 2008, pp. 51-57

Abstract: This paper presents a wireless sensor system developed for environment monitoring. Both live streaming video (1Mbyte/s) and data sensor signals with a high sampling rate (10Mbytes/s) are required to be transmitted from a mobile station to a monitoring PC in real-time. In order to obtain a 12Mbytes/s wireless data sampling rate, an embedded computer links a video camera and a Cypress signal acquisition board through USB 2.0 ports on the sensor side, whilst IEEE 802.11g is used for wireless communication of sensor signals to the monitoring PC. It is significant in wireless sensor system that the communication link should be reliable for high data throughput. The design and implementation of wireless communication for reliable and real-time sensing applications using the commercial offthe- shelf techniques are reported and discussed in this paper. The factors affecting the performance of data transmission such as the data rate, time delay, packets loss impacted by distance are experimentally investigated through the developed prototype system.

Keywords: data acquisition, Linux based, USB device operation, wireless sensor.


sai: cosiwn.2008.06.171

A Pseudo-GA for Estimating Network Traffic Matrices

Hong Tang and Tong-Liang Fan

Communications of SIWN, Vol. 4, June 2008, pp. 58-63

Abstract: A Traffic Matrix (TM) giving the traffic volume between origin and destination nodes in a network has tremendously potential utility for network capacity planning and management. Yet a major challenge remains towards achieving the full potential of traffic matrices. Many pioneer methods assumed the independence among the pairs of origins and destinations in a network, but they are inaccurate. What’s more, they didn’t consider the relation between the TM and the links. In this paper, the assignment model is developed. In order to make the problem less under-constrained, the key links are analyzed and the relation between TMs is investigated. A new method for practical and rapid inference of traffic matrices in IP networks from link count measurements and routing configuration is proposed. Finally, the accuracy of the algorithm is evaluated.

Keywords: traffic matrix estimation; assignment model; genetic algorithm


sai: cosiwn.2008.06.172

Kinetic Parameter Mining of Swarm Behavior Based on Rough Set

Yu Wu, Kai Zhou, Jie Su and Hong Tang

Communications of SIWN, Vol. 4, June 2008, pp. 64-69

Abstract: Emergent intelligence behavior in complexly adaptive system is a hot research field, among which swarm is one of the most typical models. Although some researchers studied on dynamic behaviors of swarms by using control theory, biology theory and some other interdisciplinary methods, their results still lacked quantitative conclusion or knowledge interpretation. In this paper, data mining based on rough set is introduced to study the kinetic parameters of swarm behaviors as well as given swarm model. Firstly, decision table is constructed from swarm kinetic parameters. Then, rule knowledge is achieved via discretization, attribute reduction and value reduction. It expresses the relationship between kinetic parameters and the whole emergent pattern of swarm. Consequently, these kinetic parameters are extended to a range via data mining, and the swarm model is getting more flexible. Simulation under the relevant platform has improved the creditability of the achieved knowledge.

Keywords: emergent, complexly adaptive system, swarm, rough set, kinetic parameters.


sai: cosiwn.2008.06.173

A Location-Based Service Using Triangulation with RSSI on a Wireless Sensor Network for the Guidance of the Mobility Handicapped in Subway Stations

Min-hwan Ok and Duck-shin Park

Communications of SIWN, Vol. 4, June 2008, pp. 70-76

Abstract: This work involves the pursuit of a solution to the inconveniences experienced by disabled or older persons when moving through subway stations. A location positioning technique was proposed to build a system that monitors such mobility handicapped individuals. Every node is connected ad hoc, and linked to a monitoring server. Guided information could be offered to such a mobility handicapped individual based on their monitored position information. The monitoring resolution was determined by the gaps between the deployed nodes.

Keywords: location-based service, ubiquitous sensor network, mobility handicapped, triangulation, received signal strength indication.


sai: cosiwn.2008.06.174

Predicting Consumer Acceptance of Mobile Advertising in China

Xiang Shen and Huaping Chen

Communications of SIWN, Vol. 4, June 2008, pp. 77-81

Abstract: China has the largest mobile phone subscribers group in the world and mobile advertising industry achieved a rapid development. But there is little empirical study about this market. Based on the theory of planned behavior, this paper builds an integrated research model to understand the acceptance of mobile advertising from consumer perspectives. A survey is developed and the structural equitation modeling technique is used to examine the research model. The result supports the research model in predicting consumer acceptance of mobile advertising.

Keywords: mobile advertising, mobile commerce, acceptance, theory of planned behavior, technology acceptance model


sai: cosiwn.2008.06.175

Multi Parameter Based Vertical Handoff Decision in Next Generation Networks

Anita Singhrova and Nupur Prakash

Communications of SIWN, Vol. 4, June 2008, pp. 82-87

Abstract: Traditionally, Vertical Handoff Decision Algorithm (VHDA) is based on Received Signal Strength. However, this paper aims to propose multiple parameter based algorithm, where the weight vector decides selection of base stations during Handoff. The calculation of weight vector is most critical part of the vertical handoff decision algorithm implementation. The existing method of determining weight vector by standard deviation is computationally expensive and therefore, a supervised learning based Neuro-Fuzzy approach is proposed for calculation of weight vector. The weight training using neural network is an important feature of proposed VHDA. This paper discusses the Neuro-Fuzzy technique in which fuzzy membership metrics and fuzzy based inference engine is used along with neural network for the weight adjustment in order to improve the overall efficiency of seamless mobility in next generation networks.

Keywords: handoff, mobility management, neuro-fuzzy paradigm, vertical handover decision.


sai: cosiwn.2008.06.176

PPFAS: A Port-Pair based Fair Aggregation and Scheduling in SUPANET

Ji Li, Huaxin Zeng and Hao Luo

Communications of SIWN, Vol. 4, June 2008, pp. 88-94

Abstract: Aggregate scheduling can be used to solve the scalability problem in core networks. To meet the multirequirement of Single-layer User-data Switching Platform Architecture Network (SUPANET) on QoS guarantee and high-speed switching, we propose a feasible solution called Port-Pair based Fair Aggregation and Scheduling (PPFAS) which regard all micro-flows transferred through a same portpair in switches as one macro-flow. PPFAS combine CIOQ queuing method with TRSFS scheduling mechanism to build a full distributed CIOQ switch which realizes the independent scheduling of each port. By mathematic analysis and simulations, we show the PPFAS solution having the following features: (1) The buffer demand for each Virtual Input Queue (VIQ) in the switch can be calculated quantitatively, and it is no more than 2N+1 cells at most, where N is the switch size; (2) The switch delivers the best QoS performance under the modest conditions - unbalance weight is no more than 0.5 and the workload is not above 0.9.

Keywords: aggregate scheduling, QoS, SUPANET, TRSFS, PPFAS.


sai: cosiwn.2008.06.177

Integrated Public Transport Monitoring and Information System Using Location Based Services Applications

Norleyza Jailani, Mohd Norfaizi Mihsany, Marini Abu Bakar, Noor Faezah Mohd Yatim, Salha Abdullah

Communications of SIWN, Vol. 4, June 2008, pp. 95-100

Abstract: Mobile communication system has been widely accepted amongst Malaysian users. One of the applications that are becoming popular is Location Based Services (LBS). Location data provided by mobile devices has the potential to expand information services by adding a location dimension. This is particularly true in public transportation information services. In Malaysia, commuters are becoming more dependent on public transports instead of using private vehicles due to the escalating prices of petrol and tolls, and to avoid getting trapped in traffic jams. This paper proposes the architecture of a location-based service for public transportation system which offers information on bus location to mobile users. The system consists of a back-end location data reporting system for gathering real-time position data of the buses and trains, as well as a complete front-end system targeted for mobile commuters and the transport company administration for updating information and monitoring vehicles. Global Positioning System (GPS) devices are used to determine vehicles location data which is then transmitted via Automatic Packet Reporting System (APRS). Commuters may request new updates on transport information service and find places of interest within close vicinity of bus or train stations from their mobile phones.

Keywords: location-based service, mobile application, public transportation


sai: cosiwn.2008.06.178

On the Rebroadcast Probability of an Enhanced Counter-Based Broadcast Scheme for Mobile Ad Hoc Networks

Aminu Mohammed, Mohamed Ould-Khaoua, and Lewis M. Mackenzie

Communications of SIWN, Vol. 4, June 2008, pp. 101-105

Abstract: The simplest and commonly used mechanism for broadcasting in mobile ad hoc networks (MANETs) is flooding, where each node retransmits every uniquely received message exactly once. Despite its simplicity, it can result in high redundant retransmission, contention and collision in the network, a phenomenon referred to as broadcast storm problem. Several probabilistic approaches have been proposed to mitigate this problem inherent with flooding. However, majority of these schemes uses fixed rebroadcast probability which is quite unlikely to be optimal in other simulation set up. In this paper, we propose a rebroadcast probability function which takes into account key simulation parameters like network topology size, transmission range and number of nodes in the network in determining appropriate rebroadcast probability for a given node. Simulation results reveal that this simple adaptation achieves superior performance in terms of saved-rebroadcast, number of retransmission node and end to end delay, without sacrificing reachability in dense network.

Keywords: broadcasting, MANETs, flooding, broadcast storm problem, rebroadcast probability, saved rebroadcast, reachability.


sai: cosiwn.2008.06.179

An Efficient Rate Control Scheme for Mobile Video Streaming

Peng Lin, Yanbing Liu and Yang Zou

Communications of SIWN, Vol. 4, June 2008, pp. 106-109

Abstract: Mobile video streaming is a technique that the receiver could continuously view the video content while receiving the data. An interactive framework of mobile video streaming is presented in a practical way. To reduce the effects caused by the fluctuation of wireless network, an adaptive rate control method is proposed for priority process streaming with QoS control. The transmitting data rate adapts to a suitable level according to the network status. The bandwidth is dynamically allocated according to the network status, and the client buffer occupancy and playback requirement. The transmission rate is determined by the optimal data rate control algorithm that can obtain the maximal utilization of the client buffer and minimal occupation of the network bandwidth. Simulation scenario has been created to help analyse the behaviour of adaptive streaming over wireless channel. Both cases with and without the proposed methods were examined and the result shows the proposed model is able to improve the overall performance in cases when the channel conditions fluctuate.

Keywords: wireless, adaptive streaming, rate control


sai: cosiwn.2008.06.180

Using Contracts for Self-Management

Jens Bruhn and Guido Wirtz

Communications of SIWN, Vol. 4, June 2008, pp. 110-115

Abstract: Enterprise Applications (EA) represent a family of highly complex software systems for supporting the business of companies. For their development and initial configuration, the concept of Component Orientation (CO) provides a sound foundation, leading to software architectures consisting of functionally decoupled components. These are loosely connected through contractually specified interfaces. To facilitate the administration of EAs, the vision of Autonomic Computing (AC) is based on the idea to assign low level management tasks to the system itself. In this paper we present a new approach for self-management based on contracts of interfaces. During runtime, contract adherence is evaluated. For violations, different reactions can be configured, reaching from forwarding erroneous calls over the initiation of rollbacks for confining the effects of already processed requests up to the destruction of affected elements. Our platform specific approach addresses EAs based on the Enterprise Java Beans standard (EJB), version 3.0.

Keywords: enterprise applications, autonomic computing, self-management, design by contract


sai: cosiwn.2008.06.181

Adaptive Coalition Structure Generation in Cooperative Multi-agent Systems

Giovanni Rossi and Gabriele D'Angelo

Communications of SIWN, Vol. 4, June 2008, pp. 116-122

Abstract: In multiagent systems a coalition structure is a collection of pair-wise disjoint subsets of agents whose union yields the entire population. Given a characteristic function quantifying the worth of agent subsets, searching for optimal coalition structures (i.e. where the sum of subsets’ worth is maximal) is a well-known NP-hard combinatorial optimization problem. While existing algorithms (either deterministic or stochastic) deal with time-invariant goal functions, the focus here is on dynamic settings, where the worth of agent subsets possibly varies over time in an unknown and unpredictable fashion. The aim is to design an adaptive dynamic process generating coalition structures with high worth most of the times. To this end, detecting variations in the worth of agent subsets becomes crucial. The proposed method takes into account such (possible) changes by intensifying the exploration activity whenever they are detected. The performance with respect to the worth of optimal coalition structures is evaluated through simulations.

Keywords: Adaptive Coalition Structure Generation, Coalitional Game, Simulation, Dynamic and Non-superadditive Environment, Cooperative Multiagent System.


sai: cosiwn.2008.06.182

Self-Organised Service Network for Dynamic Logistics: Framework and Case Study

Alexander Smirnov, Nikolay Shilov, Tatiana Levashova and Alexey Kashevnik

Communications of SIWN, Vol. 4, June 2008, pp. 123-127

Abstract: The objectives of the research presented in the paper are to provide a means for efficient logistics management in decentralised environment taking into account ever-changing environment of logistics activities. The research addresses these objectives proposing a contextaware decision support system (DSS) for dynamic logistics. The DSS is implemented as a set of Web-services. The paper proposes a methodology describing main principles of building such DSSs and DSS architecture. Real-time modelling of logistics situations and solving problems of dynamic logistics are carried out by a set of Web-services, which self-organise a service network according to the context. The architecture presents a set of Web-services making up the DSS. The methodology is demonstrated by an example of solving routing and transportation planning problems applied to disaster relief actions.

Keywords: self-organised service networks, web-services, service-oriented architecture, context-aware decision support, dynamic logistics, disaster relief


sai: cosiwn.2008.06.183

A Self-organizing Algorithm for Mobile Agents Formation Control

Miguel A. Sánchez-Acevedo, Ernesto López-Mellado and Félix Ramos-Corchado

Communications of SIWN, Vol. 4, June 2008, pp. 128-132

Abstract: This paper addresses the problem of establishing and maintaining structured formations of physical mobile agents using self-organization strategies. Formations are achieved through information obtained from local interactions and, if it is necessary, the formation is balanced when new elements are integrated to the group; furthermore the position of mobile agents is not defined a priori. A Petri net based methodology for obtaining the algorithm is proposed. Simulations were performed in wedge formation to validate the proposed approach.

Keywords: mobile agent formation, self-organization strategies, Petri nets


sai: cosiwn.2008.06.184

Digital Semiochemical Coordination

Holger Kasinger, Jörg Denzinger and Bernhard Bauer

Communications of SIWN, Vol. 4, June 2008, pp. 133-139

Abstract: Indirect interactions by olfactory stimuli between living organisms are a powerful mechanism for self-organizing coordination in biology. Various adoptions of this paradigm for computer systems however are mainly based on the usage of digital pheromones, although these chemical substances are only one type that mediate indirect interactions. Biology provides an ingenious diversity of such substances, all grouped by the term semiochemicals. In this paper we adopt the principles behind semiochemical coordination in biology and present a model that defines a coarse-grained architecture of selforganizing computer systems based on indirect interactions. This model allows for any combination of semiochemical coordination mechanisms within one single system architecture, which will pave the way for an easier engineering of selforganizing solutions better adapted to complex problems. We further demonstrate how to efficiently combine different types of semiochemical coordination into one mechanism, based on pollination in biology, and evaluate its application to instances of pickup and delivery problems.

Keywords: biologically-inspired, decentralized coordination, engineering self-organizing emergent systems.


sai: cosiwn.2008.06.185

On the Encapsulation and Reuse of Decentralized Coordination Mechanisms: A Layered Architecture and Design Implications

Jan Sudeikat; and Wolfgang Renz

Communications of SIWN, Vol. 4, June 2008, pp. 140-146

Abstract: The effective and reliable coordination of agent activities is a momentous problem for Multi–Agent System (MAS) developers. Particularly challenging is the decentralized coordination of agents that enables systems to exhibit self–organization. Natural phenomena typically serve as design metaphors and developers apply Decentralized Coordination Mechanisms (DCMs) that have been inferred from biological, physical or social systems. This paper addresses the utilization of DCMs as reusable software components.Current development practices give little guidance for DCM selection and force developers to manually design, implement and tune mechanism parameters ad hoc, leading to highly specialized algorithms. Here, we propose a layered software architecture that encapsulates DCMs in (multiple) coordination spaces. A generic, annotation–based interface allows to separate agent coordination from agent functionality, therefore enabling DCM reuse and facilitating application (re)designs, i.e. mechanisms exchange and parameter adjustments. Implications for development procedures are discussed and the application of the layered architecture is exemplified in a resource allocation case study.

Keywords: decentralized coordination, coordination mechanism

reuse, feedback, multi-agent system, self–organization.


sai: cosiwn.2008.06.186

Global Order From a Minimal Local Resource Allocation Strategy

Mariusz Jacyno, Seth Bullock, Terry Payne and Michael Luck

Communications of SIWN, Vol. 4, June 2008, pp. 147-152

Abstract: In this paper we explore the relationship between local and global behaviour in a simple model of utility computing infrastructure as the system heterogeneity, load and reliability are varied. To do this, we implement minimally complex agent strategies for which we can identify the fundamental generic feedback underlying system behaviour. Such feedback must be balanced by any utility computing infrastructure if decentralised control is to become an effective technique for preserving stable functionality.

Keywords: decentralised control, multi-agent system, selforganisation, decentralised resource allocation


sai: cosiwn.2008.06.187

A Social Reinforcement Teaching Approach to Social Rules

Jose Fabregat-Pinilla, C. Carrascosa and Vicente J. Botti

Communications of SIWN, Vol. 4, June 2008, pp. 153-157

Abstract: Reinforcement Learning has been an active research field both in the intelligent agents area and game theory for many years. The application to multi-agent systems is more recent but very promising. A social setting has inherent problems for a reinforcement learning agent, such as the allocation of reward between the agents or the convergence to a policy in a non-stationary setting. However, it has also much to offer to the multi-agent community. A social setting as an environment to be learnt is characterized by the agents that inhabit it. A learning agent can learn how to interact and collaborate with other agents or which rules regulate the social setting. In this paper explicit reinforcement teaching is proposed as a way to teach the set of social rules to a new agent in a social environment. Social reinforcement between agents, combined with trust, are used as a way to learn social rules or to emerge them from the interaction. A case study has been developed to test the validity of the presented approach, teaching a set of social rules to new agents that become then teachers themselves.

Keywords: multi-agent systems, social learning, reinforcement learning


sai: cosiwn.2008.06.188

Autonomic Computing: Exploring Self-Organization and Self-Configuration Using Models of Excitable Media

Luis M. Fernández-Carrasco, Hugo Terashima-Maríın, and Manuel Valenzuela-Rendón

Communications of SIWN, Vol. 4, June 2008, pp. 158-162

Abstract: This research paper explores the idea to use models of excitable media in order to get two characteristics autonomic computing systems must have: self-organization and self-configuration. Excitable media is capable of modeling complex coherent actions using simple rules; thus, reducing the difficulty and effort that, so far, the design of autonomic systems has shown. In order to delve into this idea, an excitable environment that holds autonomous agents was built, and a study of how exciting this world affects the behavior of these agents was conducted. The obtained results open the doors to a new way to design autonomic systems, more specifically, autonomic systems modeled after excitable media.

Keywords: autonomic computing, self-organization, agentbased modeling, excitable media


sai: cosiwn.2008.06.189

Multi-layer Coordinated Adaptation Based on Graph Refinement for Cooperative Activities

Ismael Bouassida Rodriguez, Nicolas Van Wambeke, Khalil Drira, Christophe Chassot, andMohamed Jmaiel

Communications of SIWN, Vol. 4, June 2008, pp. 163-167

Abstract: Future network environments are likely to be used by cooperative applications. Indeed, the recent advent of peerto- peer systems motivates this assumption. In this paper, we present a method that relies on graphs and graph grammar productions to automatically refine a high level service interactions representation of an activity into a deployment topology at the Middleware and the Transport level. At both levels, algorithms for service optimization are presented. The different models and algorithms are implemented in a case study of CMS-like operations for crisis management.

Keywords: dynamic re-configuration, self-adaptation, graphgrammar, context awareness, cooperative activities


sai: cosiwn.2008.06.190

Management-by-Exception - A Modern Approach to Managing Self-Organizing Systems

René Schumann, Andreas D. Lattner and Ingo J. Timm

Communications of SIWN, Vol. 4, June 2008, pp. 168-172

Abstract: A well established design technique for software systems is to define their application context and its behavior in all possible scenarios. Those systems are typically deterministic in terms of their reaction to inputs from their environment. A disadvantage of those systems is their lack of flexibility. They fail when they have to work in a context not regarded at design time. It was shown that multiagent systems can offer such a flexibility. But from an external position their behavior cannot always be anticipated, which in fact is a major obstacle using this technology in commercial applications. Therefore, we propose an approach that combines the advantages of deterministic behavior and flexibility. We use a multiagent system with an internal management structure. Using the management-by-exception approach allows a maximum of local decision freedom and therefore of flexibility, while ensuring an acceptable overall system performance.

Keywords: strategic management, multi-agent system, selforganization, management-by-exception


sai: cosiwn.2008.06.191

A Service-Driven Approach for Sensor Web Progressive Self-Organization

Nafaâ Jabeur and Youssef Iraqi

Communications of SIWN, Vol. 4, June 2008, pp. 173-177

Abstract: Several works have addressed the issue of selforganization in distributed systems such as sensor webs (SWs). However, straightforward SW self-organization is still extremely difficult. Moreover, SWs are moving toward supporting multiple services. In this context, we tackle the problem of self-organization using a service-driven approach. Since several services can be deployed simultaneously, the SW can be seen as the overlap of active Service-Oriented Subnets (SOSs). Each SOS results from a self-organizing effort during which relevant sensors collaborate for the provision of a particular service. In the proposed architecture, the SW is modeled as a Multi Dynamic Spatial Information Grid (MDSIG), where every DSIG, implemented by a specific SOS, is an infrastructure that has the capability of providing a service on-demand. Using the MDSIG and SOS concepts, we propose an innovative Progressive Self-Organization (PSO) approach to implement SW self-organization.

Keywords: dynamic spatial information grid, sensor web, service-oriented subnet, progressive self-organization.


sai: cosiwn.2008.06.192

Model Driven Autonomic Software Development Support

Y. Abuseta and A. Taleb-Bendiab

Communications of SIWN, Vol. 4, June 2008, pp. 178-182

Abstract: Despite the rapid advancements in autonomic systems research and development their design and engineering support is still under active research. This paper presents our early findings towards the development of a proposed Model Driven Development (MDD) method for autonomic software. This method distinguishes itself from other similar MDD methods, in that it provides support starting from the upstream stage of intention (goal or business process) definition to the downstream stages of code generation. This has the advantage over others to providing transformations and mappings starting from the business process layer to the service and component layers. The paper starts with a brief review of model-driven development for autonomic systems. This is followed by a detailed description of the proposed method, which is illustrated with a simplified proof of concept development example. The paper concludes with general critical review of the method including proposed further work.

Keywords: autonomic systems, domains, MDD, service, tasks.


sai: cosiwn.2008.06.193

Inter-Operating Grid Workload Framework with Node Autonomy

Wei Liu, José Cunha, Vitor Duarte and Tiejian Luo

Communications of SIWN, Vol. 4, June 2008, pp. 183-188

Abstract: The Grid provides us a way to utilize distributed and heterogeneous resources to benefit the next generation of scientific research. While many grids are currently supporting the work of many scientists, the approach on how to sense a complex grid system is still under developing. Considering the explosively growing amount of data during monitoring and the dynamic context of the grid computing environment, we cannot expect the monitor to fulfill its responsibility in a independent, centralized way, without diverse, flexible and scalable access capability. To address this problem, this paper proposes an interoperable workload model towards developing a Grid Workload Community (GWC). It decouples the workload module from the complex grid environment and makes the interconnected performance traces configurable, self-adaptive, thus reusable to other grid systems. With this design, we can discover and share the grid workload in a more timely way among the federated grid nodes. This model eases the process to collect, analyze, transfer and use the grid performance data and guarantees the grid workloads are managed in a decentralized interoperable way.

Keywords: grid; workload modeling; performance benchmark.


sai: cosiwn.2008.06.194

Influence of Defined Parameters in Stability of Class Number in Unsupervised Learning

B. Tighiouart

Communications of SIWN, Vol. 4, June 2008, pp. 189-193

Abstract: The electric activity of the heart indicated by the appearance, on the electrocardiogram tracing (ECG), of wave called P, Q, R, S and T. The wave demarcation is a very important phase of the signal analysis. Besides it constitutes a previous step to the parameters extraction. The precision of calculated parameters then of the diagnosis depends on that. Most of the ECG wave boundaries detection algorithms are based on the matching of a one-dimensional detection function against a standard template computed from an expert controlled reference data set. Because of the variety of ECG, in this paper, we propose to enhance the method by first stratifying the shapes of the detection functions in the vicinity of the waveform boundaries into K shape specific classes by means of a Kohonen self organizing neural network. However accuracy detection of boundary of ECG depends on determinate classes. Stability of class number is based on the learning parameters of classification algorithm. In this work we show the influence of learning parameters values on the quality of determined classes (or learning process).

Keywords: Kohonen map, self organizing, signal ECG, unsupervised learning.






Copyright © 2008 Systemics and Informatics World Network