Catálogo de publicaciones - libros

Mobile agents are increasingly used in various Internet-based applications such as electronic commerce, network management, and information retrieval. If mobile agents can dynamically add, delete, and change their functionalities at run-time, they can sufficiently satisfy characteristics such as scalability, dynamicity, robustness, and performance, which are important in Internet-based applications. In this paper, we introduce Dynamic Composition of Functionality (DCF) based on code mobility, which enables mobile agents to dynamically compose their functionalities at run-time. In order to realize DCF based on code mobility. we propose necessary language constructs and implement a platform for scalable mobile agents, called DC-AOP. We also present a case study using DC-AOP to show the usefulness of our proposal.

This paper presents first a formal development methodology that enables a specifier to add complexity progressively into the system design, and to formally validate each step wrt client’s requirements. Second, the paper describes the application of this methodology to agent-based systems, as well as development guidelines that help the specifier during the development of such systems. The methodology and the development guidelines are presented through an agent market place example.

The meta-learning problem has become an important issue in the recent years. This has been caused by the growing role of datamining applications in the global information systems of big companies which want to obtain benefits from the analysis of its data. It is necessary to obtain faithfull application rules that guide the datamining process in order to achieve the best possible models that explain the databases. We follow an inductive approach to discover these kind of rules. This paper explains the MAS-based information system we use for mining and meta-learning, and how the scalability problem is solved in order to support a community of many software agents.

Scalability is a many-sided property which can be captured in a scalability metric that balances cost, volume, timeliness and other attirbutes of value in the system, as a function of its size. Studies of typical metrics can reveal which parts of the agent infrastructure are most critical for scalability. Simple metrics are investigated for systems dominated by agent behaviour. As a system is scaled up, the length of the average tour increases and this has a major effect on performance and scalability limits. Senstivity experiments show that infrastructure improvements can improve scalability but they will not alter the general conclusions.

There is an increasing demand for designers and developers to construct ever larger multi-agent systems. Such systems will be composed of hundreds or even thousands of autonomous agents. Moreover, in open and dynamic environments, the number of agents in the system at any one time will fluctuate significantly. To cope with these twin issues of scalability and variable numbers, we hypothesize that multi-agent systems need to be both (able to determine the most appropriate organizational structure for the system by themselves at run-time) and (able to change this structure as their environment changes). To evaluate this hypothesis we have implemented sucham ulti-agent system and have applied it to the domain of automated trading. Preliminary results supporting the first part of this hypothesis are presented: adaption and self-organization do indeed make the system better able to cope with large numbers of agents.

In this statement, we sketch how to employ mobile agents for distributed computing and how such a solution compares to traditional approaches.

One of the reasons for attending to agent technology is the evergrowing complexity of information systems and the increasing difficulty to foresee and plan for all potentially arising situations. Unfortunately, some pressing issues in practical applications still remain outside the focus of agent research. Chief among them is robustness in environments that are prone to disturbances, failures or uncontrolled interactions. In our research effort, we provide a middleware, based on database transactions, that formally guarantees robustness of execution of agent actions and automates many standard actions carried out in case of disturbances. We built a simulator to test the scalability of the proposed middleware. The simulator also gives us a better understanding of the behavior of the various planning strategies reflected by the transaction trees executed by the middleware and is a valuable tool for evaluating the performance of any multi-agent system before its actual deployment. In this paper, the agent transaction model executed by the middleware is outlined. The simulator is described and the experimental results of simulating a growing number of antagonist agents with collective conflicts over a common database are presented.

The KRAFT project has defined a generic agent-based architecture to support — the process of locating and extracting knowledge from multiple, heterogeneous on-line sources, and transforming it so that the union of the knowledge can be applied in problem-solving. KRAFT focuses on knowledge in the form of expressed against an object data model defined by a shared ontology. KRAFT employs three kinds of agent: locate appropriate on-line sources of knowledge; transform heterogeneous knowledge to a homogeneous constraint interchange format; fuse the constraints together with associated data to form a dynamically-composed constraint satisfaction problem, which is then passed to an existing constraint solver engine to compute solutions. The KRAFT architecture has been designed to be scalable to large numbers of agents; this paper describes the features of the architecture designed to support scalability.

The aim is to study the behaviour (in terms of utility) of a large number of autonomous agents under various environmental situations. Our experimental framework is an extension of the Swarm simulator with an array of intelligent agents developed in C++ and Prolog.

Numerous agencies and agent systems have been developed as vehicles to deliver novel types of e-commerce services to users. However service agents in one agency are often unable to interoperate or co-operate with agents from another vendor’s agency. Clearly, standardization in this area would help to create a more ubiquitous market and supporting infrastructure for agent-based services. We analyze the relevant standards for agents, highlighting the FIPA (Foundation for Intelligent Physical Agents) agent standards. Standard specifications ought to be grounded within a practical framework, which provides a reference implementation for developers and users. We also highlight the importance of open agent systems and open source implementations and their role in seeding the market for agent technology.