Catálogo de publicaciones - libros

Services provide an universal basis for the integration of applications and processes that are distributed among entities, both within an organization and across organizational borders: This paper presents a model-driven approach to design interoperable agents in service-oriented architectures (SOA). The approach provides a foundation for how to incorporate autonomous agents into a SOA using principles of model-driven development (MDD). It presents a metamodel (AgentMM) for a BDI-agent architecture and relates AgentMM to a platform-independent model for SOAs (PIM4SOA). In this paper we mainly concentrate our discussions on the service and process aspects of SOA and how transformations to agent technology would look like. We argue that this mapping allows the design of generic agent systems in the context of SOAs that are executable in an adaptive and flexible manner.

In this paper negotiation is presented as a solution to the formation of virtual organization in domains with many parties having (partially) unknown constraints and profiles and in which the environment is dynamic by nature. The solution presented is based on the MAGNET negotiation system, for which an extension is presented, that allows for last minute changes and failure management. An efficient algorithm is presented for supplier agents, incorporating preferences, and other constraints related to existing individual plans). Combining the algorithms for supplier agents, with a simple customer agent specification, and the ability to iterate the bidding, MAGNET is extended to deal with domains as described above. A case study in logistics using real data from a logistics company shows the validity of the approach.

Continuations are a well established programming concept, allowing to explicitly capture and resume the current program state. They are present in several functional programming languages (such as Scheme), in concurrent models (such as process calculi or Hewitt actor model), and more recently in dynamic programming languages (such as Ruby, Smalltalk, Python, and even Javascript or Java). They have been applied to automaton programming, cooperative threads, compilation techniques, and have lastly raised interest in web application programming. This paper shows how this concept happens to be especially useful and elegant to program agent behaviors (or behavioral components), while increasing code readability and ease of writing. The proposed approach especially facilitates modular interaction protocol implementation, one of the main difficulties in conversational agents engineering.

Agent mobility requires additional security standards. While the theoretical aspects of mobile agent security have been widely studied, there are few studies about the security standards of current agent platforms. In this paper, test cases are proposed to assess agent platform security. These tests focus on malicious agents trying to attack other agents or the agency. Currently, they have been carried out for two agent platforms: JADE and SeMoA. These tests show which of the known theoretical security problems are relevant in practice. Furthermore, they reveal how these problems were addressed by the respective platform and what security flaws are present.

This paper proposes a new trust and reputation model to assist decision making process into agents in P2P environments, taking WSMO as the base for definition of tasks to contract. This work shows the integration of trust and reputation model and WSMO in two ways: 1) how agents use WSMO as ontology to define their requirements, responses, domain-dependent features and metrics; and 2) how the Web services discovery process in WSMO may be improved using trust and reputation criteria given by the model from data stored by consumer agents in previous interactions.