Jiewen Luo

Agent-based grid computing

The grid concept has recently emerged as a vision of future network based computing, by enabling seamless integration of computing systems and clusters, data storage, specialized networks and sophisticated analysis and visualization software. Intelligent agents can play an important role in helping achieve the grid vision. In this paper we propose a model for agent-based grid computing from the implementation point of view. Based on MAGE, a multi-agent framework, we present a reference implementation of this agent-based grid computing model called AGEGC. We believe that AGEGC will be a useful platform for research on agent-based grid computing.

Distributed data mining on agent grid: issues, platform and development toolkit

Centralized data mining techniques are widely used today for the analysis of large corporate and scientific data stored in databases. However, industry, science, and commerce fields often need to analyze very large datasets maintained over geographically distributed sites by using the computational power of distributed systems. The Grid can play a significant role in providing an effective computational infrastructure support for this kind of data mining. Similarly, the advent of multi-agent systems has brought us a new paradigm for the development of complex distributed applications. During the past decades, there have been several models and systems proposed to apply agent technology building distributed data mining (DDM). Through a combination of these two techniques, we investigated the critical issues to build DDM on Grid infrastructure and design an Agent Grid Intelligent Platform as a testbed. We also implement an integrated toolkit VAStudio for quickly developing agent-based DDM applications and compare its function with other systems.

Multi-agent Cooperation: A Description Logic View

In this paper we propose dynamic description logic for formalizing multi-agent cooperation process with a clearly defined syntax and semantics. By combining the features of knowledge representation and reasoning of description logic and action theory for multi-agent interaction, our logic is effective and significant both for static and dynamic environment. On the static side, we employ description logic for the representation and reasoning of beliefs and goals. On the dynamic side, we adopt the object-oriented method to describe actions. The description of each action is composed of models, preconditions and effects. It can reflect the real changes of the world and is very suitable for belief revision and action planning. Based on our logic, we investigate how to form joint goal for multi-agent cooperation. In particular, we propose an effective dynamic planning algorithm for scheduling sub goals, which is greatly crucial for coordinating multi-agent behaviors.

Internet Intelligent Platform-AGrIP

In information age, both governments and enterprises need intelligent platform to provide dynamic, real-time intelligent information processing service. This paper describes an Agent Grid based Intelligent Platform AGrIP to support service layer information processing. AGrIP integrates a comprehensive set of tools to compose a whole platform that will enable straightforward management of large-scale information from Internet and then apply data mining, case base reasoning and expert system to aid practical enterprise applications.

AGrIP: an agent grid intelligent platform for distributed system integration

Grid and agent communities both develop concepts and mechanisms for open distributed systems, albeit from different perspectives. How to apply the agent technology in Grid infrastructure to facilitate distributed system integration (virtual organization formation) becomes an interest topic in recent years. In this paper, we investigate the issue and present an agent grid intelligent platform, called AGrIP, from the implementation point of view. AGrIP is based on the FIPA-compliant multi-agent environment MAGE and has four different layers, which apply the agent interface services for interaction and communication. It can integrate legacy systems and enables interoperability between distributed heterogeneous systems (organizations), expedite dissemination and retrieval of data, automate operator tasks and support the decision-making process. Up to date, it has been applied to several practice projects as the robust infrastructure. We will expound its recent application in Agent Grid Based City Emergency Inter-Act Project as an illustration.

A multi-agent negotiation model applied in multi-objective optimization

Although both multi-objective optimization and agent technology gained a lot of interest during the last decade, many aspects of their functionality still remain open. This paper proposes the multi-agent negotiation model applied in multi-objective optimization. There are three types of agents in the system. The plan agent plans the global best benefit; the action agent plans the best benefit of the single objective; and the resource agent manages the common resource. The agents compete and cooperate to reach the global best benefit through their negotiation. The model is applied in evolutionary multi-objective optimization to realize its parallel and distributed computation, and the experiment on MAGE shows the model is effective.

An Interval-Based Knowledge Model and Query Language for Temporal Information

Time plays a crucial role in describing dynamic domain and resources on the Web. Most of current ontology languages, without notion of time or change, can only describe domain knowledge in a static sense. In order to bring structure to time-varying information, we propose an interval-based knowledge model (IKM), which is composed of a temporal knowledge schema, temporal integrity constraints and an interpretation structure. In the IKM, a group of expressive temporal constructs are provided to capture the temporal features of concepts, attributes and relationships; a group of rules are defined to enrich expressivity of the model; formal meaning is given to the model via an interpretation with model-theoretical semantics. A high-level query language is defined on the model, too.

A DDL based formal policy representation

This paper introduces the policy uniform concept model and defines the general policy specification language-GPSL based on the formal foundation of DDL. Compared with other policy languages, GPSL describes the general concept model providing the support for the policies including action, goal, utility and so on. Moreover GPSL provides the support for complex action composition and description to define complex system behavior. To resolve the policy conflicts we use meta-policy to illustrate the policy relationships and owl to define policy ontology. Finally this paper analyzes the policy implementation in autonomic computing system.

Eliminate redundancy in parallel search: a multi-agent coordination approach

Web spider is a widely used approach to obtain information for search engines. As the size of the Web grows, it becomes a natural choice to parallelize the spiders crawling process. However, parallel execution often causes redundant web pages to occupy vast storing space. How to solve this problem becomes a significant issue for the design of next generation web spiders. In this paper, we employ the method from multi-agent coordination to design a parallel spider model and implement it on the multi-agent platform MAGE. Through the control of central facilitator agent, spiders can coordinate each other to avoid redundant pages in the web page search process. Experiment results demonstrate that it is very effective to improve the collection efficiency and can eliminate redundant pages with a tiny efficiency cost.

Parallel web spiders for cooperative information gathering

Web spider is a widely used approach to obtain information for search engines. As the size of the Web grows, it becomes a natural choice to parallelize the spider’s crawling process. This paper presents a parallel web spider model based on multi-agent system for cooperative information gathering. It uses the dynamic assignment mechanism to wipe off redundant web pages caused by parallelization. Experiments show that the parallel spider is effective to improve the information gathering performance within an acceptable interaction efficiency cost for controlling. This approach provides a novel perspective for the next generation advanced search engine.