SIWN Abstracts Index
Editorial for Special Issue Section on Emergent Intelligence of Networked Agents
Guest Editor: Akira Namatame
International Transactions on Systems Science and Applications, Vol. 3, No. 1, May 2007, pp. 1-2
Emergent Intelligence on Network: A survey
Satoshi Kurihara and Takashi Ishikawa
International Transactions on Systems Science and Applications, Vol. 3, No. 1, May 2007, pp. 3-10
Abstract: The First Workshop on Emergence Intelligence on Network (WEIN’05) was held Oct. 3-5, 2005, in Ito city, famous hot spring village of Japan. This workshop concerned with emergence of intelligent behaviors over networked agents and fostering the formation of an active multi-disciplinary community on Multi-agent systems and Complex Networks. In this paper, we show the abstract of several papers presented in the workshop and the outlook for future of this community.
Keywords: adaptation, complex network, emergence intelligence, evolution, web dynamics, multi-agent, scale-free network, small-world network
Integrating Segregated Markets
Dan Ladley and Seth Bullock
International Transactions on Systems Science and Applications, Vol. 3, No. 1, May 2007, pp. 11-18
Abstract: In many cases real markets are segregated to some extent by constraints on who is readily able to trade and communicate with whom. This kind of segregation is modelled within a market constrained by an underlying network topology. The impact of segregation on market convergence is quantified, and the extent to which it is redressed by a broadcast mechanism intended to mimic the presence of information sources that are widely consulted, but imperfect, and slow to react to market change is explored.
Keywords: agent based economics, entropy, market, network.
Analysis on Public Transport Networks of Railway, Subway and Waterbus in Japan from the Viewpoint of Complex Network
Takahiro Majima, Keiki Takadama and Mitujiro Katuhara
International Transactions on Systems Science and Applications, Vol. 3, No. 1, May 2007, pp. 19-26
Abstract: Characteristics of network in the real world have attracted a number of scientists and engineers. Various findings are given from recent studies on the real world network, sometimes called complex network. The properties on the complex network are revealed mainly by early studies focused on the un-weighted relational network, in which there is no weight on links or vertices. Meanwhile, studies investigating weighted network has begun to appear in recent years. In these papers the weight denotes distance or traffic amount between vertices. The knowledge of complex network seems to provide useful information to design and construct the transport networks. It is fascinating if an algorithm developed with the knowledge of complex network is able to provide effective and optimized public transport system to reduce traffic congestions and increase redundancy of transport system under disaster circumstances. Toward this goal, this paper starts by investigating five transport networks, one railway, three subways and one hypothetical waterbus lines in Japan and their combinations from the viewpoint of complex network. Furthermore the role of waterbus network in Tokyo is made clear using measures in terms of complex network.
Keywords: complex network, efficiency, public transport system, traffic amount, weighted network.
Optimizing of Growing Queuing Network with Slight Reconstruction
Yusuke Matsumura, Hidenori Kawamura and Azuma Ohuchi
International Transactions on Systems Science and Applications, Vol. 3, No. 1, May 2007, pp. 27-34
Abstract: The high-quality service provided by large-scale telecommunication networks constructed on the Internet such as Gnutella is thought to reflect a certain feature network structure. The authors previously tried to obtain sub-optimal structure with genetic algorithms (GA) and verified that GA causes the structure to be optimal. However, since optimization with GA requires huge calculation time, using it to optimize structure in real-time is difficult. In this paper, we added a reconstruction restriction to GA that considered the effectiveness of reconstruction used by P2P protocols. As a result, the authors clarified that under some conditions, GA with small reconstruction can generate structures with acceptable waiting time.
Keywords: complex networks, network design, queuing network, peer-to-peer network, queuing system, telecommunication network.
Network Dynamics of Emergent Flocking Behavior
Masaru Aoyagi and Akira Namatame
International Transactions on Systems Science and Applications, Vol. 3, No. 1, May 2007, pp. 35-43
Abstract: Flocking is the phenomenon in which self-driven individuals, using limited information and simple rules, organize into an ordered motion. This ordering occurs at the dynamic level where ordering is not possible at static equilibrium of systems. This paper explores computational and theoretical efforts undertaken to explain emergent phenomenon among autonomous mobile agents. A class of local control laws is investigated for a collection of agents that result in global order their motion in convergence of their movement to a common while avoiding collision with obstacles. These are made possible through local control action by exploiting the graph theoretic properties of the underlying agent networks. Algebraic connectivity affects the performance and robustness properties of the overall flocking behavior. It is shown how the stability of the flocking motion is directly associated with the connectivity properties of the agent networks.
Keywords: boids, emergent, flocking behavior, network.
Editorial for the Special Issue Section on Supporting Technologies for Autonomic Computing: Methodologies, Frameworks and Tools
Guest Editors: Fábio Luciano Verdi, Marcelo Perazolo and Eduardo Nicola Ferraz Zagari
International Transactions on Systems Science and Applications, Vol. 3, No. 1, May 2007, pp. 44-44
Achieving Autonomic Web Service Integration: A Quality of Service Policy-Based Approach
Diego Zuquim Guimarães Garcia and Maria Beatriz Felgar de Toledo
International Transactions on Systems Science and Applications, Vol. 3, No. 1, May 2007, pp. 45-63
Abstract: Web service technology offers facilities for business process automation. However, Business Process Management Systems demand additional features. For instance, there is no Quality of Service (QoS) management support in the current Web service model. Consumers may require services with particular non-functional characteristics and expect quality level guarantees. The Web Services Description Language (WSDL) and Universal Description Discovery & Integration (UDDI) standards support the specification, publication and discovery of Web services based only on functional aspects. The goal of this paper is to propose a policy-based Web service infrastructure, which offers autonomic service integration using QoS attributes. The infrastructure includes brokers to select services and monitors to verify QoS. The main contributions of this paper are extensions to the Web Services Policy Framework (WS-Policy) and UDDI standards.
Keywords: autonomic service integration, business process management system, policy, quality of service, semantic web, universal description discovery & integration, web service, web services policy framework.
FOCALE – A Novel Autonomic Networking Architecture
John C. Strassner, Nazim Agoulmine and Elyes Lehtihet
International Transactions on Systems Science and Applications, Vol. 3, No. 1, May 2007, pp. 64-79
Abstract: Network resources will always be heterogeneous, and thus have different functionalities and programming models. This adversely affects interoperability. This approach establishes a common “lingua franca” through the combination of information models and knowledge engineering, which together can be used to discover and program semantically similar functionality for heterogeneous devices regardless of the data and language used by each device. This paper introduces FOCALE, a semantically rich, context-aware, policy based architecture for orchestrating the behavior of heterogeneous and distributed computing resources. This architecture is applied to Beyond 3G Networks and to Motorola’s Seamless Mobility as case studies.
Keywords: autonomic architecture, autonomic networking, FOCALE, seamless mobility.
A Symptom Visualization Framework for Autonomic Problem Determination
Marcelo Perazolo and Abdi Salahshour
International Transactions on Systems Science and Applications, Vol. 3, No. 1, May 2007, pp. 80-88
Abstract: This paper describes novel aspects of event visualization and trust establishment utilizing concepts of autonomic computing, such as symptoms, to represent, detect, evaluate and resolve incidents and problems related to the IT infrastructure management and operations. In addition, symptom visualization and processing methods are suggested in order to enable efficient pro-active avoidance of these incidents and problems before they manifest themselves. This paper describes a complete implementation of these concepts and relates real experiences in the art of event, incident and problem management.
Keywords: autonomic computing, event management, problem determination, symptoms, trust establishment.
Editorial for Special Issue on New tendencies on Web Services and Multi-agent Systems
Guest Editors: Pilar Herrero and Gonzalo Méndez
A Web Service Specification to Manage Awareness in Collaborative Grid Environments
Pilar Herrero, José Luis Bosque, Manuel Salvadores and María S. Pérez
International Transactions on Systems Science and Applications, Vol. 3, No. 2, July 2007, pp. 90-104
Abstract: In this paper, a new extension and reinterpretation of one of the most successful models of awareness in Computer Supported Cooperative Work (CSCW), called the Spatial Model of Interaction (SMI), which manage awareness of interaction through a set of key concepts, to manage task delivery in collaborative distributed systems, is presented. This model, called AMBLE (Awareness Model for Balancing the Load in Collaborative Grid Environments), also applies some theoretical principles and theories of multi-agents systems to create a collaborative and cooperative environment that can be able to provide an autonomous, efficient and independent management of the amount of resources available in a Grid environment. WS-AMBLE provides an open interface having the ability of managing different levels of awareness, allowing different Virtual Organizations to share computational resources based on open protocols and interfaces. As far as it is known, none of the last WS specifications offers functionalities useful enough as to create awareness models and none of the last WS specifications offers specific functionalities to manage task balancing delivery in collaborative grid environments.
Keywords: Collaborative environments, grid computing, load balancing, web services.
Relaxed-Criteria Negotiation for Grid Resource Allocation
Kwang Mong Sim and Ka Fung Ng
International Transactions on Systems Science and Applications, Vol. 3, No. 2, July 2007, pp. 105-117
Abstract: “G-Commerce” refers to computational economies for controlling the resource allocation in computational Grid environments. Whereas many extant G-commerce models focus on auction and commodity markets, this work focuses on designing models for Grid resource negotiation (G-negotiation). The novel feature of this work is that G-negotiation agents representing resource providers and consumers are programmed to slightly relax their bargaining criteria under intense pressure with the hope of enhancing their chance of successfully acquiring resources. Two sets of fuzzy rules representing the relaxation criteria for providers and consumers respectively are developed. Augmented with a fuzzy decision controller, the agents follow a relaxed-criteria G-negotiation protocol. The relaxed-criteria G-negotiation protocol was implemented in an agent-based testbed using Java 1.5. Empirical results obtained from simulations carried out using the testbed suggest that by slightly relaxing their bargaining terms under intense negotiation pressure, agents (i) achieved higher success rates in negotiation and (ii) required fewer negotiation rounds.
Keywords: automated negotiation, G-commerce, Grid resource management.
Toward Trust and Reputation Based Web Service Selection: A Survey
Yao Wang and Julita Vassileva
International Transactions on Systems Science and Applications, Vol. 3, No. 2, July 2007, pp. 118-132
Abstract: A trust and reputation mechanism is a mechanism using consumers’ feedbacks to identify good services from bad ones. Compared with other approaches, it has more advantages in solving the selection problem for web services. The paper proposes a typology to classify trust and reputation systems using the three criteria, centralized or decentralized, person or resource, global or personalized. Inspired by the criteria, some potential research directions for web service selection are pointed out.
Keywords: reputation, trust, Web service selection.
Recommending Web Services via an Agent Federation
M. Brian Blake and Michael F. Nowlan
International Transactions on Systems Science and Applications, Vol. 3, No. 2, July 2007, pp. 133-144
Abstract: As service-oriented computing (SOC) gains popularity and large numbers of services become available over the Internet, a manual (i.e., on-demand) search and discovery approach to obtaining candidate web services may not be feasible. Consumers will require the ability to have services recommended to them as they become available. In a SOC environment, recommender systems will be characteristically different than traditional human user-based systems. Recommendations must be made in accordance with the standard operations within a service-oriented architecture (SOA) whereas machine-to-machine interactions are prevalent. This paper introduces the notion of an agent-mediated recommender system that uses enhanced syntactical approaches for the discovery of relevant services. A further innovation is the use of multiple agents as middleware to collaborate in determining the most effective recommendation while implicitly federating web service registries.
Keywords: intelligent agents, recommendation systems, web service discovery.
Integration of Web Services in an Agent Oriented Methodology
Rubén Fuentes-Fernández, Iván García-Magariño, Jorge J. Gómez-Sanz and Juan Pavón
International Transactions on Systems Science and Applications, Vol. 3, No. 2, July 2007, pp. 145-161
Abstract: This paper introduces a view of web services following the Web Services Architecture. This architecture considers web services as abstractions implemented with special programs called agents. The paper discusses the nature of these agents from the agent research perspective. Aiming at reusing the experience of agent research, the paper proposes an integration of concepts from the Web Services Architecture within an existing agent oriented software engineering methodology, the INGENIAS methodology.
Keywords: agent, agent oriented software engineering, meta-modeling, web services.
Banking Frauds: An Agent-Based Training Framework to Follow up the Swindlers Learning Process
Mauricio Paletta, Pilar Herrero and Gonzalo Méndez Pozo
International Transactions on Systems Science and Applications, Vol. 3, No. 2, July 2007, pp. 162-168
Abstract: Electronic bank transactions are very common today. Services given by an Automatic Teller Machine (ATM), for example, are very popular and widely used by bank clients. Unfortunately, in the same way as the use of these devices is increasing, the proliferation of different frauds to try to violate these systems to steal user’s money is also increasing. In this paper, a framework designed to follow up the swindlers’ agents learning process is presented. This framework is based on an open and flexible architecture that emphasizes on the swindler agents learning process to fulfil not only more human-like agent behaviour but also a more realistic interaction with the environment.
Keywords: ATM, electronic bank fraud, IVA, learning, architecture.
A Multi-Agent System for Auction-Based Resource Discovery in Semantic-Aware BC Mobile Commerce
Vedran Podobnik, Krunoslav Trzec and Gordan Jezic
International Transactions on Systems Science and Applications, Vol. 3, No. 2, July 2007, pp. 169-182
Abstract: Next-generation mobile networks will create semantic-aware environments populated with diverse types of ubiquitous communication-enabled devices which will provide a remarkable selection of heterogeneous resources (information, multimedia, services and products). Consequently, such an environment will require efficient mechanisms which can match demands (requested resources) to supplies (available resources), anywhere and anytime. In this paper, an economic approach to solving this problem combined with AI concepts is proposed. A multi-agent system for enabling B2C mobile commerce within the semantic-aware environment of next-generation mobile networks is designed. The proposed multi-agent system provides an appropriate solution for both mobile consumers (who often posses several mobile terminals) and businesses that act as resource providers. The efficiency of our model is realized by utilizing mobile software agents and by applying a resource discovery process which uses two-level filtration of available resources before a final ranked set of eligible resources is recommended to consumers in response to their needs. The filtration processes not only consider the semantic information associated with available resources, but also analyze ratings regarding the actual performance of resource providers (with respect to both price and quality) and the prices paid by providers for advertising their resources.
Keywords: intelligent software agent, semantic web services, resource discovery, ubiquitous computing, mobile commerce.
RADIC - Towards a General Method for Integrating Reactive and Deliberative Layers
James Kramer and Matthias Scheutz
International Transactions on Systems Science and Applications, Vol. 3, No. 3, October 2007, pp. 183-192
Abstract: Hybrid architectures have been developed to preserve the responsiveness of reactive layers while also providing the benefits of higher-level deliberative capabilities. The challenge of hybrid architecture design is to integrate layers that operate on different time scales, use different notions of spatial relations, maintain varying degrees of state, etc. Techniques for integration, however, are typically very specific to the application or particular hybrid architecture.
A general method is proposed for constructing hybrid architectures via the integration of pre-existing reactive and deliberate layers that requires minimal modifications. Its potential is demonstrated in navigation experiments on a robot, achieving performance comparable to a commonly used hybrid system with specially designed layers.
Keywords: hybrid architectures, navigation, obstacle avoidance, path planning, reactive deliberative integration.
Andrew Cooper and Andrew Martin
International Transactions on Systems Science and Applications, Vol. 3, No. 3, October 2007, pp. 193-202
Abstract: Security concerns currently deter or prohibit many organisations from leveraging the benefits of the grid. When sensitive data is placed under the control of third-party infrastructure it is difficult to obtain assurances that it will be appropriately protected.
The paper describes a novel grid platform architecture based on a secure root of trust. This component provides a secure environment for grid job execution that resists attack even if the host itself is compromised. Trusted Computing, a security technology currently being integrated into an increasing number of mainstream PCs, is used for dynamic trust establishment. These elements are combined to create a novel and practical solution for the grid malicious host problem, ensuring that data integrity and confidentiality is appropriately protected for jobs that span multiple administrative domains.
In existing approaches to grid delegation privileged user credentials are given to large, complex middleware software stacks where they can be stolen and used in a user masquerade attack. The proposed solution protects delegation credentials from theft using the hardware security co-processor. Access to delegation credentials can be restricted according to a global delegation policy. The policy enables the user to specify a trusted subgrid ensuring that a large, distributed computation can never escape the trusted boundary, even if the grid job itself is compromised.
Keywords: trusted computing, grid computing, distributed systems, grid security, virtual machines, delegation.
Design of Decentralized Capturing Behavior by Multiple Mobile Robots
Yuichi Kobayashi, Kyouji Otsubo and Shigeyuki Hosoe
International Transactions on Systems Science and Applications, Vol. 3, No. 3, October 2007, pp. 203-210
Abstract: This paper discusses the design of decentralized capturing behavior by multiple mobile robots. The design is based on a gradient descent method with local information. The task of capturing a target is divided into two subtasks, enclosing task and grasping task. Analysis on convergence of the local control policy in enclosing task is given, while ensuring that the neighborhood relation of the robot system is preserved. In grasping task, the force-closure condition in decentralized form is utilized for designing a local objective function. A local estimation of the shape of the object is introduced so that each robot can decide how to move based only on local information. The proposed local control policies were evaluated by simulations and the flexibility of the system was verified to be brought by the decentralized nature of the system.
Keywords: decentralized control, multiple mobile robots, cooperative capturing, force-closure.
Translating UML Diagrams into Maude Formal Specifications
Farid Mokhati, Mourad Badri and Patrice Gagnon
International Transactions on Systems Science and Applications, Vol. 3, No. 3, October 2007, pp. 211-219
Abstract: This paper presents a formal framework supporting the translation of UML diagrams into a formal specification based on the Maude language. The proposed approach considers both static and dynamic features of object-oriented systems. We consider, in particular, UML class, state and communication diagrams. The formal and object-oriented language Maude, based on rewriting logic, supports formal specification and programming of concurrent systems. The major motivations of this work are: (1) to bind together the UML notation and the Maude language, (2) to improve the description of object-oriented systems preserving their coherence and supporting their verification and validation process. The generated Maude specifications are validated by simulation. The approach is illustrated using a concrete example.
Keywords: formal specification, Maude, object-oriented systems, UML, translation, verification and validation.
On Automatic Parameter Configuration Mechanism for Data Transfer Protocol GridFTP
Takeshi Ito, Hiroyuki Ohsaki and Makoto Imase
International Transactions on Systems Science and Applications, Vol. 3, No. 3, October 2007, pp. 220-226
Abstract: In this paper, an automatic parameter configuration mechanism for GridFTP is proposed. This mechanism optimizes the number of parallel TCP connections by utilizing the analytic result of GridFTP throughput. The proposed mechanism first measures the network status (e.g., the goodput and the round-trip time of GridFTP data channels) at the GridFTP client. Based on these measurement results, it adjusts the number of parallel TCP connections for maximizing the GridFTP goodput. Three operational modes, MI (Multiplicative Increase), MI+ (Multiplicative Increase Plus), and AIMD (Additive Increase and Multiplicative Decrease) are proposed in this paper, each of which takes a different strategy for adjusting the number of parallel TCP connections. Performance of the proposed automatic parameter configuration mechanism is evaluated through simulation experiments. This paper reveals that the proposed automatic parameter configuration mechanism significantly improves the performance of GridFTP.
Keywords: automatic parameter configuration mechanism, Grid computing, GridFTP, parallel TCP connections
Collaborative Multi-Robot Monte Carlo Localization in Assistant Robots
R. Barea, E. López, L.M. Bergasa, S. Álvarez and M. Ocaña
International Transactions on Systems Science and Applications, Vol. 3, No. 3, October 20073, pp. 227-237
Abstract: This paper presents an algorithm for collaborative mobile robot localization based on probabilistic methods (Monte Carlo localization) used in assistant robots. When a root detects another in the same environment, a probabilistic method is used to synchronize each robot’s belief. As a result, the robots localize themselves faster and maintain higher accuracy. The technique has been implemented and tested using a virtual environment capable to simulate several robots and using two real mobile robots equipped with cameras and laser range-finders for detecting other robots. The result obtained in simulation and with real robots show improvements in localization speed and accuracy when compared to conventional single-robot localization.
Keywords: mobile robots, collaborative multi-robot localization, Monte Carlo localization, assistant robots.
Reinforcing Design Rationale in Software Projects Developed in Academic Environment
Débora M. B. Paiva, André P. Freire, Daniel Lucrédio, Rosana T. V. Braga and Renata P. M. Fortes
International Transactions on Systems Science and Applications, Vol. 3, No. 3, October 2007, pp. 238-248
Abstract: During a software development process, the recording and management of information related to design rationale (description of “how” and “why” an artifact has been designed in a certain way) are important issues to achieve better support for software maintenance. In this paper, a model to represent Design Rationale in Software projects developed in Academic Environment, named DR-SAE, is introduced, presenting its requirements, architecture and implementation. Experiments comparing the DR-SAE model with traditional approaches were carried out. In general, it was observed an improvement on motivation for design rationale registering when the DR-SAE model was used.
Keywords: design rationale, software process, and academic research.
Identifying Roles, Goals and Scenarios at the Early Requirement Phase of Multiagent System Modeling
Nancy Ruiz, Adriana Giret and Vicente Botti
International Transactions on Systems Science and Applications, Vol. 3, No. 3, October 2007, pp. 249-256
Abstract: The Early Requirements specification for a Multiagent System (MAS) is a crucial phase before applying any methodology to develop a MAS. In this paper, it is presented some early requirement specification guidelines for MAS development. These guidelines are based on the Decomposition Theory of complex systems. The guidelines will help the software engineer to build an initial MAS structure defined by roles, goals, and role relations. Moreover, the approach can be used to define an early requirement phase in any MAS methodology.
Keywords: early requirements, modeling, multiagent systems.
Machine Learning for Performance Predictions on Space-shared Computing Environments
International Transactions on Systems Science and Applications, Vol. 3, No. 3, October 2007, pp. 257-268
Abstract: We present a study of machine learning techniques for performance predictions on space-shared computing resources. By examining the problem in a general background of knowledge discovery and learning from data, new performance metrics such as effective capacity can be introduced. A set of new attributes is defined to characterize the resource states and the corresponding distance functions are formulated, which makes it possible for the predictions of queue wait time and effective capacity. Local learning is adopted as the framework, and a set of machine learning techniques including genetic algorithms and M-Tree search are applied to improve predictions. We also investigate parameter tuning for subsets of training data and develop a novel algorithm for adaptively selecting the tuning methods. All these components are combined to provide a new and practical approach towards performance predictions. Experimental studies are conducted to evaluate the proposed solution using real workloads from production systems.
Keywords: Machine Learning, Performance Predictions, Space-shared computing resources, Grid Computing.
Using Multi-Agent Systems to Study Cells Interaction
Lynda Dib and Zahia Guessoum
International Transactions on Systems Science and Applications, Vol. 3, No. 3, October 2007, pp. 269-278
Abstract: This paper presents our multi-agent framework (named CellInvasion) for modelling and predicting the emergent behaviour resulting from the presence of distinct environmental conditions that lead to bad interaction of cells in their tissue. As the migration of tumoural cells through the biologic gates constitutes the major dynamic event of tumoural invasion, the objective of CellInvasion is to be a virtual world of cellular biology while simulating the cellular behaviour specially the migratory behaviour in order to help its specialists to better understand, to good interpret and to warn changes of cell states according to its actual internal state and to the state of its environment. The Architecture of CellInvasion is composed of a set of reactive agents. These agents use an ontology (named OntoCell) to communicate and to represent their knowledge. OntoCell describes the knowledge of cellular biology. CellInvasion is implemented with a multi-agent platform (named DIMA).
Keywords: ontology, multi-agent systems, simulation, cells, cellular links, migration, tumour, cancer, invasion.
Throughput Performance of the BEA WebLogic JMS Server
Robert Henjes, Michael Menth and Valentin Himmler
International Transactions on Systems Science and Applications, Vol. 3, No. 3, October 2007, pp. 279-288
Abstract: Communication among distributed software components according to the publish/subscribe principle is facilitated by the Java messaging service (JMS). JMS can be used as a message routing platform, if the subscribers install filter rules on the JMS server. However, it is not clear whether its message throughput is sufficient to support large-scale systems. This paper investigates the capacity of the high performance JMS server implementation Bea WebLogic. In contrast to other studies, the focus of this paper is on the message throughput in the presence of filters and show that filtering reduces the performance significantly. Also a model is presented for the message processing time at the server and validate it by measurements. This model takes the number of installed filters and the replication grade of the messages into account, and predicts the overall message throughput for specific application scenarios.
Keywords: Brokering Middleware, Service Oriented Architecture, Benchmarking and Assessment
Editorial to Special Section on Multi-Agent Systems and Large-Scale Distributed Systems
Dana Petcu and Przemysław Stpiczynski
International Transactions on Systems Science and Applications, Vol. 3, No. 4, January 2008, pp. 289-289
Analysis of Execution Anomalies in the Parallel Kohonen SOM Algorithm
Jan Kwiatkowski, Marcin Pawlik and Dariusz Konieczny
International Transactions on Systems Science and Applications, Vol. 3, No. 4, January 2008, pp. 290-295
Abstract: The parallel runtime depends not only on the utilized algorithm but also on the hardware and software environment properties. Execution time anomalies resulting from the hardware and software environment behavior or inexact performance model are well known to every scientist analyzing the performance of the parallel program. An abnormal behavior can usually be observed only for specific input data, execution parameters or hardware configurations. In the paper different sources of anomalies are discussed and the way how the performance analysis can be carried out in their presence is presented using the example of a parallelized version of the Kohonen SOM algorithm.
Keywords: parallel processing, execution anomalies, Kohonen algorithm.
Utilizing Agent Teams in Grid Resource Brokering
M. Dominiak, M. Ganzha, M. Gawinecki, W. Kuranowski, M. Paprzycki, S. Margenov and I. Lirkov
International Transactions on Systems Science and Applications, Vol. 3, No. 4, January 2008, pp. 296-306
Abstract: Recently it was suggested that (mobile) software agents can provide an infrastructure for resource brokering and management in Grids. In this paper we introduce a novel approach based on agent teams. Since the yellow pages-based approach was selected to facilitate resource discovery, we summarize results of our experiments to find an efficient ways of implementing yellow page service in an agent-based system. We also discuss how agents can find a team that will execute their job.
Keywords: market-based approach, multi-agent system, coordination strategies.
Distributed SQL in Spatial Data Warehouse and Software Agents Environment
Marcin Gorawski and Ewa Płuciennik
International Transactions on Systems Science and Applications, Vol. 3, No. 4, January 2008, pp. 307-313
Abstract: The article presents an environment for a distributed SQL query execution in a telemetric data warehouse. A parallel, spatial data warehouse stores information extracted remotely from meters supervised by nodes according to its geographical position. This data warehouse is a distributed data structure. The proposed software agents’ environment enables querying this structure as a local one so distribution is transparent to the user. The query evaluation consists of an analysis, decomposition, local execution and merging results. There is no need to send data between nodes. Some formal operators for executing query modification and merging sub-results are defined. The test results of SQL query realizations in different agent environments are also presented.
Keywords: Distributed data warehouse, software agents, SQL.
Optimizing Forecast of Reputation of Mobile Agents in Open Environment
Mieczyslaw A. Klopotek and Michal Wolski
International Transactions on Systems Science and Applications, Vol. 3, No. 4, January 2008, pp. 314-318
Abstract: This paper presents the design of an Open Mobile Agent Simulator. We describe in terms of UML the behavior and structure of the simulator. We compare two reputation metrics (eBay and BetaSystem) for mobile agents traveling through a network of host nodes and we point at a very important problem: the speed of recognition of the overall reputation of a group of host nodes. To solve this problem we offer an innovative technique based on forecasting trust value. We present a meta-algorithm – an optimization formula, which can be useful for many trust metrics.
Keywords: mobile agents, optimization, reputation.
An Implementation of the JADE-based A-Team Environment
Dariusz Barbucha, Ireneusz Czarnowski, Piotr Jedrzejowicz, Ewa Ratajczak-Ropel and Izabela Wierzbowska
International Transactions on Systems Science and Applications, Vol. 3, No. 4, January 2008, pp. 319-328
Abstract: The paper proposes a JADE-based A-Team environment (in short: JABAT) as a middleware supporting the construction of dedicated A-Team architectures that can be used for solving a variety of computationally hard optimization problems. The paper contains a short review of the A-Team concept, overview of JABAT functionality and a description of how this functionality can be extended to solving new problems. Use of JABAT is illustrated with a few example optimization problems solved using the proposed approach.
Keywords: JABAT, A-Team, optimization, computionally hard problems, multi-agents systems.
Abstractions of Multi-Agent Systems
Constantin Enea and Catalin Dima
International Transactions on Systems Science and Applications, Vol. 3, No. 4, January 2008, pp. 329-337
Abstract: In this paper, abstraction techniques for reasoning about knowledge in multi-agent systems are proposed. The logic used is the temporal logic of knowledge under Kleene’s 3-valued interpretation along infinite and maximal finite paths. Various weak, very weak and error preservation results are provided.
Keywords: verification, abstraction, multi-agent system, temporal logic of knowledge.
KBRMAS - Knowledge Based Grid Resource Management for Compromised Alternate Solution
P. Muthuchelvi, G. S. Anandha Mala and V. Ramachandran
International Transactions on Systems Science and Applications, Vol. 3, No. 4, January 2008, pp. 338-345
Abstract: Grid computing offers many scientific and commercial benefits, notwithstanding its technical and organizational constraints. In such a scenario, providing an efficient mechanism for allocation and management of resources is essential for a computational grid. As resource providers and requesters are independent identities, negotiation and co-ordination among them is necessary. Towards this end, the system of Knowledge Based Grid Resource Management for Compromised Alternate Solution (KBRMAS) is proposed to enhance the success rate of the agent by providing alternate solution by relaxing some of the criteria. This offers unlimited scope for judicious management of resources to the advantage of resource providers and requesters in ensuring speedy execution of processes. The proposed system offers scope for devising means of cutting down delays arising out of prolonged wait for readily unavailable resources. Quality of service is improved by negotiating with the agents with reference to the user’s specification and for providing alternate solution when discovery fails with the criteria of time and cost. The knowledge based agent whose role is outlined as part of the proposed system is thus enabled to solve the resource discovery problem in an efficient and speedy manner.
Keywords: Agent based resource management system, alternate solution, discover agent, knowledge based agent, and Resource discovery.
A Performance Based Load-Balancing Algorithm for Grid Computing
Abderezak Touzene, Khaled Day and Bassel Arafeh
International Transactions on Systems Science and Applications, Vol. 3, No. 4, January 2008, pp. 346-354
Abstract: In this paper we propose hierachical architecture to design grid computing services. We also propose a performance based load-balancing algorithm for the proposed grid computing architecture, which maximize the throughput and resources utilisation. Our load-balancing algorithm takes into account both the CPU speed of the computing units and the network bandwidth of the grid system networks. We design an analytical queueing model to derive at steady-state, performance indices such as task average response time, site or worker saturation level, and resource utilization. We present an optimality study for our algorithm and experimental results are obtained to confirm our analytical findings.
Keywords: grid computing, load-balancing, performance evaluation, queuing models, resource management, steady-state.
A Trust Management Architecture for Resource Selection in Grid
P. Varalakshmi and S. Thamarai Selvi
International Transactions on Systems Science and Applications, Vol. 3, No. 4, January 2008, pp. 355-365
Abstract: Trust Management is an important issue in a grid environment where consumers and Service Providers (SPs) are distributed geographically across autonomous administrative domains. This paper proposes a trust management architecture that supports the choice of SPs based on their trust-indices maintained at intermediaries, the brokers. SPs and Consumers are distributed across these brokers, with each of these entities being associated with more than one broker. This improves redundancy of information maintained at broker’s sites, thereby improving reliability, and eases network traffic while handling consumer requests and feedbacks. Trust-indices of brokers, SPs and consumers are updated dynamically after the completion of each transaction, enables consumers to receive responses from brokers significantly quicker compared to other reputation-based-trust models where trust-indices are computed at request-time. Effectiveness of this trust model over schemes that operate without trust and that compute trust-indices at request-time is show-cased. Marked improvement is seen in job-success-rate for various percentages of malicious entities, cost-loss to consumer-community and waiting-time for handling requests. Use of broker’s and consumer’s feedback for the computation of trust-index has a positive impact on choosing a ‘trust-worthy’ SP.
Keywords: cost-loss, Grid, job-success-rate, presence-index, reputation-based, Resource selection, trust-index, waiting-time.
Processing Semantic Web Queries in Grid
Marko Niinimaki and Tapio Niemi
International Transactions on Systems Science and Applications, Vol. 3, No. 4, January 2008, pp. 366-377
Abstract: This paper describes a design and implementation for a Grid based data retrieval and query processing system for Semantic Web applications for distributing data and services in a secure manner.
With existing Grid software, several milestones have been reached concerning resource and service discovery, data replication, and running specific applications with massive amounts of distributed data. In order to describe Grid resources, services and data, ontology-based approaches have also been utilised. This kind of approach has been described as the Semantic Grid. Based on the ideas of the Semantic Grid, a framework where services and data are declared in a catalog is presented. There, a data entity is a record in the catalog and it contains a pointer (an access URL) to a data source, an ontology map file describing the structure of data, a description of the data, and a nickname. Grid security is used in accessing data and services.
Using this framework, a distributed Resource Description Framework (RDF) query system that supports ontology maps has been implemented. The architecture and user interfaces of the system are discussed and efficiency is demonstrated by measurements in different environments. Moreover, a prototype of a distributed, ontology-aware query processor is presented.
Keywords: Grid, Grid services, RDF, ontology.
Incorporating Enhanced Resource Provisioning in a Grid Resource Management System
Inderveer Chana and Seema Bawa
International Transactions on Systems Science and Applications, Vol. 3, No. 4, January 2008, pp. 378-386
Abstract: The next-generation Grids will include services, resource reservations for predictable application service quality, and dynamic adaptation to changing load and system conditions. The resource availability in Grids is generally unpredictable due to the autonomous and shared nature of the Grid resources and stochastic nature of the workload resulting in a best effort quality of service. Therefore, there is a need to develop techniques for better resource utilization mechanisms for future Grid Environments. This paper discusses the techniques for better resource provisioning in a Grid environment, which can be provided through key data generated by the base models. This data can be put to appropriate use, thus providing optimum resource usage for the Grid users. The implementation of resource provisioning has been demonstrated with the help of a prototype for a campus wide Grid.
Keywords: grid resource management, resource model, resource provisioning, web services.
Copyright © 2007-2008 Systemics and Informatics World Network