Abstract: We propose constraints as the appropriate computational constructs for the design of agents with the task of selecting, merging and managing electronic information coming from such services as Internet access, digital libraries, E-mail, or on-line information repositories. Speci cally, weintroduce the framework of Constraint-Based Knowledge Brokers, which are concurrent agents that use so-called signed feature constraints to represent partially speci ed information and can exibly cooperate in the management of distributed knowledge. We illustrate our approach by several examples, and we de ne application scenarios based on related technology such asTelescript and work ow management systems. Key Words: multiagent coordination, agent-interaction, distributed problem solving, signed feature constraints, negotiation, cooperation strategies.

Distributed case-based reasoning architectures have the potential to improve the overall performance of case-based reasoning systems.

This paper presents studies in learning a form of organizational knowledge called organizational roles in a multi-agent agent system. It attempts to demonstrate the viability and utility of self-organization in an agent-based system involving complex interactions within the agent set. We present a multi-agent parametric design system called L-TEAM where a set of heterogeneous agents learn their organizational roles in negotiated search for mutually acceptable designs. We tested the system on a steam condenser design domain and empirically demonstrated its usefulness. LTEAM produced better results than its non-learning predecessor, TEAM, which required elaborate knowledge engineering to hand-code organizational roles for its agent set. In addition, we discuss experiments with L-TEAM that highlight the importance of certain learning issues in multi-agent systems. Introduction Requirements like reusability of legacy systems and heterogeneity of agent representations lead to a number of c...

Rising operating costs and structural transformations such as resizing and globalization of companies all over the world have brought into focus the emerging discipline of knowledge management that is concerned with making knowledge pay off. Corporate memories form an important part of such knowledge management initiatives in a company. In this paper, we discuss how viewing corporate memories as distributed case libraries can benefit from existing techniques for distributed case-based reasoning for resource discovery and exploitation of previous expertise. We present two techniques developed in the context of multi-agent case-based reasoning for accessing and exploiting past experience from corporate memory resources. The first approach, called Negotiated Retrieval, deals with retrieving and assembling "case pieces" from different resources in a corporate memory to form a good overall case. The second approach, based on Federated Peer Learning, deals with two modes of cooperation called DistCBR and ColCBR that let an agent exploit the experience and expertise of peer agents to achieve a local task.

In literature, the mediator architecture has been proposed for taking information from distributed, heterogeneous, and often dynamic sources and making them work together as a whole. In this paper we propose a distributed case-based approach for the main problem of a mediator, i.e. rewriting queries according to mediator's schema. According to this approach we use a case memory as mediator's schema. Therefore, such a schema is not static (as in other systems) but is dynamically updated through the cooperation with information sources and other mediators, strongly influenced by the queries submitted by a consumer. From the analysis of different cooperation strategies arises that it is more efficient and effective for a mediator to directly cooperate with information sources, when the sources are few. Otherwise, it is more efficient to cooperate with other mediators. 1

The focus of this paper is information brokers within Information Discovery (ID). We describe Cumulative Duality matrices, an instrument to deal with design criteria for such information brokers. ID is the synthesis of Information Retrieval and Information Filtering, where information brokers act as middle-agents. There are numerous design criteria for information brokers. Since these stem from ID, they exhibit a dual nature. The duality of the criteria is shown to be cumulative. In the form of a matrix, cumulative duality can be used as a design instrument for information brokers. Contents 1 Introduction 3 2 Cumulative duality in Information Discovery 4 2.1 Duality in Information Discovery . . . . . . . . . . . . . . . . 4 2.2 Cumulative duality in directed communication . . . . . . . . 5 2.3 Cumulative Duality Matrix . . . . . . . . . . . . . . . . . . . 5 3 Instantiations of CD Matrices 6 3.1 Partial environmental knowledge . . . . . . . . . . . . . . . . 6 3.2 Privacy of inte...

We propose the use of case-based plan recognition in computer system and network management. We describe why this might be useful, how a system could be designed to do this task using cases, and how this work relates to other work in plan recognition, CBR, and software agents. Introduction The explosion in networked information resources suggests a greater need for automated methods for management of network (and other) resources and information. One approach is to use specialized software agents to perform some of this network management in response to changing conditions. An ideal agent would contribute to overall system effectiveness while acting automatically in a large, dynamic, often timeconstrained environment. Suppose an agent's task is to manage some resource, such as a printer or a file server, and part of this task involves transferring work elsewhere to avoid overloaded conditions. This management would be more effective if the agent could predict the usage level and act o...

Abstract. The problem of heterogeneous case representation poses a major obstacle to realising real-life multi-case-base reasoning (MCBR) systems. The knowledge overhead in developing and maintaining translation protocols between distributed case bases poses a serious challenge to CBR developers. In this paper, we situate CBR as a flexible problemsolving strategy that relies on several heterogeneous knowledge containers. We introduce a technique called language games to solve the interoperability issue. Our technique has two phases. The first is an eager learning phase where case bases communicate to build a shared indexing lexicon of similar cases in the distributed network. The second is the problem-solving phase where, using the distributed index, a case base can quickly consult external case bases if the local solution is insufficient. We provide a detailed description of our approach and demonstrate its effectiveness using an evaluation on a real data set from the tourism domain. 1

The work presented in this thesis deals with techniques to improve problem solving control skills of cooperative agents through machine learning. In a multi-agent system, the local problem solving control of an agent can interact in complex and intricate ways with the problem solving control of other agents. In such systems, an agent cannot make effective control decisions based purely on its local problem solving state. Effective cooperation requires that the global problem-solving state influence the local control decisions made by an agent. We call such an influence cooperative control. An agent with a purely local view of the problem solving situation cannot learn ...

this paper, I discuss two areas of work in plan recognition for HCI. The first, expectation driven plan recognition, recognizes user plans in order to communicate via a dialogue with the user. The second, case-based plan recognition, attempts to recognize what sorts of plans a user is executing by observing his or her actions, with the goal of providing assistance and managing system resources.