Catálogo de publicaciones - libros

There is a lot of confusion on what the environment of a multi-agent system (MAS) comprises. Sometimes, researchers refer to the environment as the logical entity of a MAS in which the agents and other resources are embedded. Sometimes, the notion of environment is used to refer to the software infrastructure on which the MAS is executed. Sometimes, environment even refers to the underlying hardware infrastructure on which the MAS runs.

Our research focuses on situated MASs, i.e. MASs in which agents have an explicit position in the environment. In this paper, we propose a three-layer model for situated MASs that considers agents as well as the environment as first-order abstractions. The aim of this model is to clarify the confusion between the concept of the environment and the infrastructure on which the MAS is deployed. The top layer of the model consists of the MAS application logic, the middle layer contains the software execution platform, and the physical infrastructure is located in the bottom layer. Starting from this model, we propose a classification of situated MASs based on the physical infrastructure of the MAS. We illustrate the different classes with examples from the research community and our own practice. We apply the three-layer model to each example. The models show that agents and the environment are abstractions that crosscut the three layers of the model.

In a Multi-Agent Based Simulation (MABS) special attention must go to the analysis, modeling and implementation of the environment. Environments for simulation of real world problems may be complex. Seeing the environment as a monolithic structure only reduces our capacity to handle large scale, real-world environments. In order to support this type of environments, we propose the use of an holonic perspective to represent the environment and the agents. In our approach, agents and environment are represented by holons. The environment defines a holarchy. Each agent belong to a specific holon in this holarchy following its needs.

Even if the multi-agent paradigm has been evolving for fifteen years, the development of concrete methods for problem solving remains a major challenge. This paper focuses on reactive multi-agent systems because they provide interesting properties such as adaptability and robustness. In particular, the role of the environment, which is effectively where the system computes and communicates, is studied. From this analysis a methodology to design or engineer reactive systems is introduced. Our approach is based on the representation of the problem’s constraints considered as perturbations to stabilize. Agents are then defined, in the second place, as a means of regulating the perturbations. Finally, the relevancy of our proposition is justified through the development of two solving models applied to real and complex problems.

In this paper, we discuss the model of an environment for a geographically based simulation system. The environment is structured as a graph in which nodes represent locations and edges represent paths between locations. The space is decomposed into a network of cells which are managed by cell controllers. In order to visualize location information at various levels of abstraction, we define the environment as a cell hierarchy.

The E4MAS community is leading an effort to accept environments of multi-agent systems as a first-class entity, distinguishing via the environment from the environment’s role in message transport. This paper defines classes of interaction (sequential and multi-agent, direct and indirect) and environments (physical and virtual, persistent and amnesic, dynamic and static). These notions provide an underpinning for proper acknowledgement of the roles of MAS environments and for powerful MAS design techniques that use indirect interaction. We explore the limitations of MAS models that are restricted to message passing and suggest research directions for constructing more powerful models.

Whenever a multiagent system is designed, many dependencies in the system are identified and must be solved in a correct way. Coordination deals with the management of such dependencies. For that, two complementary viewpoints can be distinguished: manages intra-agent aspects while essentially deals with inter-agent aspects. On the basis of this separation of concerns, the paper discusses the need of infrastructures for objective coordination. As in usual agent software platforms, this can be done by offering implicit support for objective coordination, by establishing the conditions necessary for running agent programs and maintaining agent interactions. Other infrastructures such as Electronic Institutions go one step further and shape the governing aspects of objective coordination. However, this is usually done through dedicated middle-agents that belong to the institution. An alternative approach is to transfer the governing or regulating responsibility from institutional agents to the of a multiagent system. A promising way of doing this is to view the environment as a rule-based infrastructure that defines reactions to events. This has the advantage of allowing for the definition of laws that not only regulate agent interaction (as most work in governed interaction), but action within the environment. We illustrate this approach by several examples in different domains of laws.

Most environments for multi-agent systems limit themselves to providing message transport and white/yellow page services. While these are generic facilities, in some domains other services are necessary, which may map real-world services provided by institutions. The Electronic Institution concept represents the virtual counterpart of real-world institutions, and one of its benefits is to provide a regulated and trustful environment by enforcing norms and providing specific institutional services. This paper presents some of such institutional services. Ontology-based services are provided to assist agent interaction, making the establishment of business agreements more efficient. After the establishment of an agreement through an appropriate negotiation process, it is necessary to verify the execution of the resulting contract. For this, we introduce an institutional normative environment based on the concept of institutional reality and norms.

Overhearing has been proposed recently as a model of indirect interactions in Multi–Agent Systems. Overhearer agents receive messages that were not primarily sent to them, as when someone hears a conversation among others. Overhearing has been modeled essentially as message broadcasting, but this approach raises several issues of scalability and appropriateness of the mental state of overheard agents.

In this paper, we motivate and propose a model of overhearing that copes with these issues by introducing an to handle overhearing. We define key notions with focus on the environment perspective, model them and their relations, and detail an algorithm that describes the environmental process for agent interactions. We finally illustrate our approach with an electronic market scenario.

While agents and environments are two intimately connected concepts, most approaches for multi-agent development focus on the agent-specific part of the system, whereas the handling of concerns related to the environment is often neglected or delegated to implementation level constructs. In this paper we demonstrate that building on an environment specification with expressive semantics is instrumental in designing agents that are capable of flexible and complex interactions. We propose a modeling approach that allows describing the concrete aspects of a multi-agent system as well as its conceptual and cognitive aspects within a single coherent conceptual framework by grounding all aspects in the environment. This framework enables an efficient development process built around the rapid prototyping and iterative refinement of multi-agent system specifications by applying model-driven design techniques to the system in its entirety.

Stigmergy (the coordination of agents through signs they make and sense in a shared environment) was originally articulated in the study of social insects. Its basic processes are much simpler than those usually used to model human-level cognition. Thus it is an attractive way to coordinate agents in engineered environments such as robotics or information processing. Stigmergic coordination is not limited to insects. Humans regularly use environmentally-mediated signals to coordinate their activities. This paper develops a schema for analyzing stigmergy among humans, discusses examples (some using a computational environment and others antedating digital computation), and suggests how the use of such mechanisms may be extended.