Detection of undesirable communication patterns in multi-agent systems

Communication design is usually complex in Multi-Agent Systems (MAS) because of dynamic emergent behaviours. The lack of proper quantitative measures to assess alternative solutions and guide an iterative development makes this design even harder. The aim of this work is to efficiently find and describe communication patterns that should be avoided in these systems and identify the agents involved in these patterns. For this purpose, this research presents a suite of novel metrics and classification rules that, respectively, measure agents’ communication and classify their results to describe patterns. This work also provides tools for automatically measuring the metrics and applying the classification rules. In order to evaluate this work, the results of applying these metrics and classification rules have been compared with the quality attribute of performance in several MAS. Performance is measured as the time between a user’s request and the MAS response, and partially represents the factor of the quality of service. The experiments gather four agent-oriented communication designs that belong to two different domains: Crisis-management and Cinema ticket selling. The study reveals that the detected communication patterns are related with performance, and that the proposed metrics can arguably guide the design of communications improving the overall performance of systems.     

An infrastructure to support cooperation of knowledge-level agents on the semantic Grid

Can a Knowledge-Level layer be located in the Semantic Grid infrastructure? Is it possible to design an Agent Communication Language (ACL) which enables Knowledge-Level agents to cooperate in a geographically distributed Semantic Grid despite nodes’ failures or malfunctions? This paper tries to address the above Semantic Grid challenges presenting an agent-based Open Service Architecture which integrates geographically distributed agents in a Semantic Grid. The architecture is well integrated with standard Internet components and technologies and supports communication among Knowledge-Level agents. The role of agents is to retrieve, execute and compose available services providing more sophisticated instances of them. Inter-agent communication is realized by exploiting an advanced Agent Communication Language which supports a fault-tolerant anonymous interaction protocol and satisfies a set of well defined Knowledge-Level programming requirements. Here, we present the design of the architecture and of the Agent Communication Language as well as their implementation. The architecture is evaluated by means of several case studies which highlight the main feature of our proposal. The main advantage of our approach is to demonstrate that different issues, such as high level inter-agent communication and fault tolerance, can be successfully integrated in Grid infrastructures which provide Web Services maintaining a clean design of the architecture and a Knowledge-Level characterization.     

Ergonomic considerations in the design of interactive languages for process control

The paper introduces the concept of software ergonomics and stresses the importance of respecting the working habits of the user while designing the software aids. Language design criteria are presented under two broad categories: those involving the syntax of the language and those involving its semantics. These criteria are then exemplified using an actual experience of designing an interactive language (TOOL) for a large public utility.     

基于移动Agent与P/S混合模式的分布对象中间件异步通信研究

针对分布对象中间件异步通信中时间、空间和流程的三刻度解耦问题进行了研究,在分析了当前分布对象中间件通信机制所存在的问题基础上,基于CORBA分布环境引入了P/S发布订阅系统,构建了基于P/S模式的CORBA通信体系.同时在PIS中间件代理体系中,引入移动Agent代理技术构建了基于Agent与P/S混合模式的异步通信算法,解决了基于中间代理的分布对象中间件异步通信的三刻度解耦问题.     

Local strategy learning in networked multi-agent team formation

Networked multi-agent systems are comprised of many autonomous yet interdependent agents situated in a virtual social network. Two examples of such systems are supply chain networks and sensor networks. A common challenge in many networked multi-agent systems is decentralized team formation among the spatially and logically extended agents. Even in cooperative multi-agent systems, efficient team formation is made difficult by the limited local information available to the individual agents. We present a model of distributed multi-agent team formation in networked multi-agent systems, describe a policy learning framework for joining teams based on local information, and give empirical results on improving team formation performance. In particular, we show that local policy learning from limited information leads to a significant increase in organizational team formation performance compared to a random policy.     

Coordination of communication in robot teams by reinforcement learning

In multi-agent systems, the study of language and communication is an active field of research. In this paper we present the application of Reinforcement Learning (RL) to the self-emergence of a common lexicon in robot teams. By modeling the vocabulary or lexicon of each agent as an association matrix or look-up table that maps the meanings (i.e. the objects encountered by the robots or the states of the environment itself) into symbols or signals we check whether it is possible for the robot team to converge in an autonomous, decentralized way to a common lexicon by means of RL, so that the communication efficiency of the entire robot team is optimal. We have conducted several experiments aimed at testing whether it is possible to converge with RL to an optimal Saussurean Communication System. We have organized our experiments alongside two main lines: first, we have investigated the effect of the team size centered on teams of moderated size in the order of 5 and 10 individuals, typical of multi-robot systems. Second, and foremost, we have also investigated the effect of the lexicon size on the convergence results. To analyze the convergence of the robot team we have defined the team’s consensus when all the robots (i.e. 100% of the population) share the same association matrix or lexicon. As a general conclusion we have shown that RL allows the convergence to lexicon consensus in a population of autonomous agents.     

The State of the Art in Agent Communication Languages

Like societies of humans, there is a need for agents in a multi-agent system to rely on one another, enlist the support of peers in order to solve complex tasks. Agents will be able to cooperate only through a meaningful communication language that can bear correctly their mental states and convey precisely the content of their messages. In search for the ideal agent communication language (ACL), several initiatives like the pioneering work of the Knowledge Sharing Effort and the Foundation for Intelligent Physical Agents (FIPA) are paving the way for a platform where all agents would be able to interact regardless of their implementation environment. ACL is a new field of study that could gain from a survey in expanding its application areas. For this purpose, we examine in this paper the state of the art in ACL design and suggest some principles for building a generalized ACL framework. We then evaluate some existing ACL models, and present the current issues in ACL research, and new perspectives. Received 22 July 1999 / Revised 23 March 2000 / Accepted 2 May 2000     

A visualisation tool to aid exploration of students’ interactions in asynchronous online communication

Much research in recent years has focused on the introduction of virtual learning environments (VLEs) to universities, documenting practice, and sharing experience ( [Armitage et?al., 2001], [Browne and Jenkins, 2003], [Moron-Garcia, 2004] and [Weller, 2007]). Attention has been directed towards the importance of online dialogue for learning as a defining feature of the VLE.Communicative tools are an important means by which VLEs have the potential to transform learning with computers from being passive and transmissive in nature, to being active and constructivist (Pavey & Garland, 2004). However, practical methods of reviewing and analysing online communication to trace cycles of real dialogue (and learning) have proved somewhat elusive. Qualitative methods are also under-used for VLE discussions, since they demand new sets of research skills for those unfamiliar with them, and can be time intensive.This paper describes a visualisation tool to aid the analysis of online communication. The tool has two purposes: first, it can be used on a day-to-day basis by teachers or forum moderators to review the development of a discussion and to support appropriate interventions. Second, the tool can support research activities since the visualisations generated provide the basis for further qualitative and quantitative analysis of online dialogue.The visualisation software is designed to encode interaction types simply and quickly. The software was tested and then used to analyse data from a sample of forums within the Moodle VLE. The paper discusses both the method of visualisation and analysis of the online interactions as a pilot for further research analysing interaction in discussion forums.     

Hohfeld in cyberspace and other applications of normative reasoning in agent technology

Two areas of importance for agents and multiagent systems are investigated: design of agent programming languages, and design of agent communication languages. The paper contributes in the above mentioned areas by demonstrating improved or novel applications for deontic logic and normative reasoning. Examples are taken from computer-supported cooperative work, and electronic commerce.