Catálogo de publicaciones - libros

This paper attempts to articulate the general role of infrastructure for multi-agent systems (MAS), and why infrastructure is a particularly critical issue if we are to increase the visibility and impact of multi-agent systems as a universal technology and solution. Second, it presents my current thinking on the socio-technical content of the needed infrastructure in four different corners of the multi-agent systems world: science, education, application, and use.

Before the powerful agent programming paradigm can be adopted in commercial or industrial settings, a complete environment, similar to that for other programming languages, must be developed. This includes editors, libraries, and an environment for the completion of agent tasks. The DECAF[] Agent architecture is a general purpose agent development platform that was designed specifically to support concurrency, distributed operations, support for high level programming paradigms, and high throughput. The architecture has been designed with built-in scalability which adapts itself to multiple processor architecture and highly distributed multi-agent systems. DECAF supports research efforts in planning and scheduling with modular design. The architecture also supports application development and has current developments in social modeling, middle agents, information extraction, and proxy operations. DECAF also supports the next step in the progression of the programming paradigm by allowing “flexible” and “structured persistent” actions []. This paper is a case study of the development of the DECAF architecture, tools that have been developed concurrently to support programming and testing, and some of the more significant applications designed using DECAF.

In this paper, we describe an infrastructure for cooperative work support in Multi-Agent Systems (MAS). The infrastructure is based on a concept of Agora which can be considered as a facilitator of cooperative work. Basic features of the Agora based system as well as some implementation details (including communication adapter, message wrapper, proxy and default agent) are presented. An example of Virtual Shopping Mall as a general framework for modeling agent-based intelligent software services in mobile communications is used for illustration of the approach.

The design and analysis of multi-agent systems is difficult due to complex agent capabilities and rich interactions among agents. Experimentation is a crucial step in gaining insight into the behavior of agents. Experiments must be flexible, easily configurable, extensible, and repeatable. This paper presents the Sensible Agent Testbed, which supports these requirements. The CORBA infrastructure of the Testbed platform and the formally-defined interfaces among Testbed components are described. The Testbed promotes many levels of modularity, facilitating parallel development of agents and agent capabilities. This approach provides many opportunities for different types of experiments. Experimental setup through a configuration file is simplified using the Init File Maker, which has the capability to automate the production of multiple configurations. Overall, the Sensible Agent Testbed provides a solid infrastructure supporting multi-agent experiments.

In this paper, we present MadKit (multi-agent development kit), a generic multi-agent platform. This toolkit is based on a organizational model. It uses concepts of groups and roles for agents to manage different agent models and multi-agent systems at the same time, while keeping a global structure.

We discuss the architecture of MadKit, based on a minimalist agent kernel decoupled from specific agency models. Basic services like distributed message passing, migration or monitoring are provided by platform agents for maximal flexibility. The componential interface model allows variations in platform appearance and classes of usage.

This paper describes an architecture for modeling cooperating systems of communicating agents. The authors’ goal is not that of providing a framework to implement multi-agents systems (there are tools—such as CORB A, Java and DCOM—that do an excellent job on that), but rather to provide an architectural metaphor upon which collaborative multi-agent systems could be modeled. The approach is based on requirements defined with a practical view of the communicational and resource-oriented nature of distributed collaborative multi-agent systems.

This paper suggests formal frameworks that can be used as the basis for defining, reasoning about, and verifying properties of agent systems. The language, Little-JIL is graphical, yet has precise mathematically defined semantics. It incorporates a wide range of semantics needed to define the subtleties of agent system behaviors. We demonstrate that the semantics of Little-JIL are sufficiently well defined to support the application of static dataflow analysis, enabling the verification of critical properties of the agent systems. This approach is inherently a top-down approach that complements bottom-up approaches to reasoning about system behavior.

Coordination, which is the process that an agent reasons about its local actions and the (anticipated) actions of others to try to ensure the community acts in a coherent fashion, is an important issue in multi-agent systems. Coordination is a complicated process that typically consists of several operations: exchanging local information; detecting interactions; deciding whether or not to coordinate; proposing, analyzing, refining and forming commitments; sharing results, and so on. We argue that facets of these different operations can be separated and bundled into two different layers.The lowerlayer pertains to and operations, i.e., the detailed analysis of candidate tasks and actions, the formation of detailed temporal/resource-specific commitments between agents, and the balancing of non-local and local problem solving activities. In contrast, the upper-layer pertains to coordination tasks such as the formation of high-level goals and objectives for the agent, and decisions about whether or not to coordinate with other agents to achieve particular goals or bring about particular objectives. Detailed domain state is used at this level to make these high-level coordination decisions. In contrast, decisions at the lower-level do not need to reason about this detailed domain state. However, reasoning about detailed models of the performance characteristics of activities, such as their temporal scope, quality, affects of resource usage on performance, is necessary at this level. In this view, the layers are interdependent activities that operate asynchronously.

With the proliferation of software agents and smart hardware devices there is a growing realization that large-scale problems can be addressed by integration of such stand-alone systems. This has led to an increasing interest in integration infrastructures that enable a heterogeneous variety of agents and humans to work together. In our work, this infrastructure has taken the form of an integration architecture called . We have deployed Teamcore to facilitate/enable collaboration between different agents and humans that differ in their capabilities, preferences, the level of autonomy they are willing to grant the integration architecture, their information requirements and performance. This paper first provides a brief overview of the Teamcore architecture and its current applications. The paper then discusses some of the research challenges we have focused on. In particular, the Teamcore architecture is based on general purpose teamwork coordination capabilities. However, it is important for this architecture to adapt to meet the needs and requirements of specific individuals. We describe the different techniques of architectural adaptation, and present initial experimental results.

The RoboCup Soccer Server and associated client code is a growing body of software infrastructure that enables a wide variety of multiagent systems research. This paper describes the current Soccer Server and the chaxmpion CMUnited soccer-playing agents, both of which are publically available and used by a growing research community. It also describes the ongoing development of FUSS, a new, flexible simulation environment for multiagent research in a variety of multiagent domains.