ArticlePDF Available

Abstracting communication in distributed agent-based systems

Authors:

Abstract and Figures

This paper discusses abstraction components for agent based communication. We report back on our experience in deploying such components for developing agent based solutions and we draw some conclusions on the main benefits and challenges of communication abstractions for multi-agent systems. We finally argue that some of these abstractions have the potential to be re-used or integrated on top of traditional (non-agent) distributed systems.
Content may be subject to copyright.
A preview of the PDF is not available
... In both its centralized and distributed formulation, the CSP framework has been extended to allow variants, which increase real-world applicability. These are in particular: (1) soft CSP, which allows to the classical notion of constraints the possibility of dealing with important features such as fuzzyness, uncertainty optimization, probability and partial satisfaction [6], and (2) dynamic CSP, where constraints may change for external reasons or depending on choices made during the search [13]. ...
... In general, there is no consensus on a commonly accepted definition of what an agent is and what are, or should be, its main properties. However, the common colloquial understanding is that the term agent indicates an entity (person, organization, system, etc.) that (1) acts on behalf of some other entity (an owner), (2) in an autonomous fashion. Thus, an agent is given the mandate to achieve defined goals. ...
... The terms "agent" and "agent communication" have been used in the DCSP and MAS worlds with slightly different meanings. While in the constraint community an agent is a computational entity acting as a decision maker following pre-defined coordination mechanisms (i.e., DCSP algorithms) and sharing an implicit common representation of the world with other agents (i.e., no explicit use of structured communication stack and ontologies), from a MAS perspective an agent is autonomously deciding whether or not to follow specific coordination mechanisms and can communicate with other agents by means of structured semantic-grounded message exchange [2]. In a way, the DCSP agent definition represents a looser version of MAS formalization, also indicated in the MAS community as "weak agency". ...
Article
Full-text available
Although constraint satisfaction techniques and agent technology have been combined in several ways, especially over the last few years, it is still debated how, beyond ad hoc approaches, these two computational paradigms can be effectively integrated. The aim of this paper is to analyze and discuss how constraint satisfaction techniques can be used to engineer agents' reasoning and resolve coordination of distributed agents, either cooperative or self-interested, by reporting on our concrete experience.
... Abstraction methods play an important role in system design as theycan be used to represent the structure of the system in an ew space so that reasoning in the newm odel is simpler than that in the original [Cal02]. Foragent communication, an abstract model had been introduced in [Hel03] as illustratedinFigure 1. ...
... At anypoint of the conversation, ac ertain type of messages is expected. The patterns of message exchange are called interaction protocol and theyrepresent the highest abstraction component in the agent communication stack [Cal02], [Bon02]. ...
... The following two papers [17,18] present a semantic communication stack that includes the message layer and the content layer among others. They emphasize in the need for agreements on the formalisms used in each layer and in the need to manage the dependencies between layers. ...
Article
Full-text available
Nowadays it is increasing the tendency of using agents on be-half of information systems in their interactions with other autonomous information systems. Moreover, there exists an actual interest on getting these interactions at a semantic level, that is, beyond the syntactic level interaction provided by the XML format standard. In this paper we present a mechanism that gives an step forward in the global goal of achieving a semantic interoperability among agents. The mechanism is based on ontological commitments on the classes of messages interchanged by agents of different information systems in an scenario where agents act in a sincere, helpful and liberal way. Furthermore those messages are considered as individuals of OWL classes, and we take advantage of the reasoning supporting OWL ontologies.
Conference Paper
Software agents are often seen as flexible goal-driven software technology suitable for dynamic environments such as business ecosystems. The inherently open and unpredictable nature of such environments, however, renders unsuitable many widely used agent design mechanisms, such as agent interaction protocols. This paper describes an approach which addresses the need for flexible inter-agent interactions. Our approach gives agents freedom of action and enhances their ability to respond to unexpected events in open environments. We base our argument on the limitations of protocol-based approaches used to formalize agent interaction. We then investigate the possibility of protocol-free agent interaction to enable flexible operation of agent systems in dynamic environments
Conference Paper
Full-text available
Like in any distributed systems, communication in Multi-Agent System (MAS) plays an important role as it helps agents sharing knowledge and facilitating cooperation. Communication in MAS may present at various forms; it can be exchanging of comprehensible messages, or requesting to perform an action, or negotiating cooperation. Agent commu-nication has common properties like in any other distributed systems. But it is the use of ACL messages that differentiates agent communication from that in other distributed systems. Based on the layered model for agent communication, there have been many studies done in order to optimize communi-cation performance between agents in MAS. In this paper, we present an evaluation of some typical Message Transport Protocol (MTP) implementations for agent communication. The performance tests are conducted on the JADE Agent Platform.
Conference Paper
Full-text available
New advances in the areas of Internet and network communications have facilitated in some way the work with distinct data systems located at different places. However, what it is still missing is the possibility of a real and efficient interoperation among those systems. In this paper we present a new mechanism that favors the interoperability of agent based data systems. That mechanism permits an agent of a system 1) to send suitable messages to agents of another system without requiring the establishment of a common communication pattern in advance; and 2) to understand, completely or partially, the messages that are sent by agents of other systems. Moreover, the mechanism also allows the communication among agents that follow different standard languages such as FIPA-ACL and KQML. The main support of the proposed mechanism is an ontology that is divided into three interrelated layers and four different categories.
Conference Paper
As applications become increasingly distributed and networks provide more and more connection facilities, applications require more and more interconnections, thus communication takes a central part of modern systems. Hence, since large applications parts have been underlined like databases systems or graphical user interface, the goal is to wonder, if we can say the same for the communication part of applications.
Article
Information technology is evolving from focusing on local systems to encompassing a more global interaction among multiple systems in enterprises and communities. On the one hand, new advances in the area of network communications have facilitated in some way the intercommunication among heterogeneous information systems located at different places. However, what is still missing is the possibility of a real and efficient interoperation among those systems in an open environment as the one favoured by Internet. On the other hand, agent technology provides platforms where cooperative work of information systems is concernable because of the software agents working on behalf of these information systems. But, nowadays that cooperation is in general restricted and requires a laborious a priori preparation. In this paper we present the features of a formal ontology that can play a relevant role in the development of a new kind of information systems interoperation frameworks. The ontology includes classes and properties for describing communication acts among agents. We claim that the communication acts ontology provides interoperability support due to the recognition of communication acts from one Agent Communication Language (acl) as instances of communication acts in another acl. Sometimes the comprehension will not be complete, but partial comprehension of the communication may be useful and preferable to the “not understood” answer given nowadays. Terms of the ontology are described as classes or properties using the Web Ontology Language owl.
Article
Full-text available
This paper focuses on the Organ Transplant Management (OTM) process defined as the set of activities that need to be coordinated before the surgery operation itself can take place. The main goal is to understand how the different OTM tasks are interrelated, in order to build an innovative software system that aims to support medical practitioners involved in the various decision-making steps. The current work and the extensive discussions with experts in this area provided ground for the definition of the OTM system. We argue that this tool, when integrated in real hospitals and transplant centres, will facilitate many of the tasks currently performed by humans, and will allow optimisation of precious resources such as time and personnel.
Conference Paper
Agents often utilise the services of other agents to perform tasks within multi agent systems. To achieve this, an agent must first locate another agent that has the capability to provide a desired service (i.e. a service provider agent), and then interact with it. To communicate with a service provider, an agent requires information about: 1) the service provider agent's interface; 2) the ontology that defines concepts used by the provider agent; and 3) the agent communication language (ACL) the agent uses so that it can parse and understand the communication. Currently deployed MASs encode the interface description and the ontology within the capability description of a service provider, but assume a common ACL between communicating agents. Middle agents support the discovery of service providers based on the advertised capabilities of the service providers. This advertisement defines the provider agent's interface, and may reference the ontology used by the provider agent. However, the requester agent still requires information about the ACL used by the provider to be able to communicate with it. This paper demonstrates how agents can communicate with each other without making assumptions about the ACLs used, by presenting a template based shallow parsing approach to message construction/decomposition, thus greatly simplifying and improving the robustness of inter-agent communication.
Article
The engineering of complex software systems is still an immature discipline beset with a continuing malaise, the symptoms of which are an inability to produce systems on time, within budget and which actually function as required. So acute is the problem with software that it has become known as 'the software crisis' and so long has it persisted that it cannot be diagnosed as immaturity alone.
Article
interpretation of PROLOG programs. In Abstract Interpretation of Declarative Languages, pages 181--198. Ellis Horwood Limited, 1987. [Men64] E. Mendelson. Introduction to Mathematical Logic. Van Nostrand Reinhold, 1964. [MH69] J. McCarthy and P. Hayes. Some philosophical problems from the standpoint of artificial intelligence. In B. Meltzer and D. Michie, editors, Machine Intelligence 4, pages 463--502. Edimburgh University Press, 1969. [MH91] I. Mozetic and C. Holzbaur. Extending EBG by abstraction operators. In Proceedings EWSL-91, Porto, Portugal, 1991. Springer-Verlag. [Moz90a] I. Mozetic. Abstractions in model-based diagnosis. In Working Notes of AAAI-90 Workshop on Automatic Generation of Approximations and Abstractions, pages 64--75. AAAI, 1990. [Moz90b] I. Mozetic. Reduction of diagnostic complexity through model abstractions. In Proc. 1990 Intnl. Workshop on Principles of Diagnoses, Stanford University, 1990. [MSW90] A.K. Modi, D.M. Steier, and A.W. Westerberg. Learning ...
Chapter
The long-anticipated revision of this best-selling book offers the most comprehensive, up-to-date introduction to the theory and practice of artificial intelligence. Intelligent Agents. Solving Problems by Searching. Informed Search Methods. Game Playing. Agents that Reason Logically. First-order Logic. Building a Knowledge Base. Inference in First-Order Logic. Logical Reasoning Systems. Practical Planning. Planning and Acting. Uncertainty. Probabilistic Reasoning Systems. Making Simple Decisions. Making Complex Decisions. Learning from Observations. Learning with Neural Networks. Reinforcement Learning. Knowledge in Learning. Agents that Communicate. Practical Communication in English. Perception. Robotics. For those interested in artificial intelligence.
Article
The relatively new field of artificial life attempts to study and understand biological life by synthesizing artificial life forms. To paraphrase Chris Langton, the founder of the field, the goal of artificial life is to “model life as it could be so as to understand life as we know it.” Artificial life is a very broad discipline which spans such diverse topics as artificial evolution, artificial ecosystems, artificial morphogenesis, molecular evolution, and many more. Langton offers a nice overview of the different research questions studied by the discipline [6]. Artificial life shares with artificial intelligence (AI) its interest in synthesizing adaptive autonomous agents. Autonomous agents are computational systems that inhabit some complex, dynamic environment, sense and act autonomously in this environment, and by doing so realize a set of goals or tasks for which they are designed.
Article
Publish/subscribe is considered one of the most important interaction styles for the explosive market of enterprise application integration. Producers publish information on a software bus and consumers subscribe to the information they want to receive from that bus. The decoupling nature of the interaction between the publishers and the subscribers is not only important for enterprise computing products but also for many emerging e-commerce and telecommunication applications. It is often claimed that object-orientation is inherently incompatible with the publish/subscribe interaction style. This flawed argument is due to the persistent confusion between object-orientation as a modeling discipline and the specific request/reply mechanism promoted by CORBA-like middleware systems. This paper describes object-oriented abstractions for publish/subscribe interaction in the form of Distributed Asynchronous Collections (DACs). DACs are general enough to capture the commonalities of various publish/subscribe styles, and flexible enough to allow the exploitation of the differences between these styles.