Increasing the Autonomy of Deliberative Agents With a Case-Based Reasoning System

Abstract

This paper shows how deliberative agents can be built by means of a case-based reasoning system. The concept of deliberative agent is introduced and the case-based reasoning model is presented. Once the advantages and disadvantages of such agents have been discussed, it will be shown how to solve some of their inconveniences, especially those related to their implementation and adaptation. The World Wide Web has emerged as one of the most popular vehicle for disseminating and sharing information through computer networks; a distributed agent-based solution for e-business, in which such agents have been used, is also presented and evaluated in this paper.

Constructing deliberative agents with case-based reasoning technology

“…This agent starts looking from a predetermined Web address and searches for new ones using natural language-processing strategies, optimized for the Web, based on the combination of the grammatical characterization of words and statistical decision techniques. 26 This agent is monitored and guided by a marketing expert. We also are studying the possibility of implementing this agent using the CBR-BDI model presented here or as proposed in Ref.…”

Agent-Based Web Engineering

“…It has been demonstrated that BDI agent implemented using a case-based reasoning system can substantially solve the two problems previously mentioned [8], [14]. Implementing agents in the form of CBR systems also facilitates their learning and adaptation.…”

“…If the proper correspondence between the three mental attitudes of BDI agents and the information manipulated by a case-based reasoning system is established, an agent with beliefs, desires, intentions and a learning capacity will be obtained. Although the relationship between agents and CBR systems has been investigated by other researchers [1], [10], [19], [26], we have proposed an approach, whose main characteristic is its direct mapping between the agent conceptualisation and its implementation, in the form of a CBR system [8], [13], [14]. Intelligent and autonomous agents may then be implemented for working on the Internet automatically, in an unsupervised way, and then may be capable of evolving without supervision.…”

“…Our proposal is also concerned with the agent's implementation and presents a "formalism" easy to implement, in which the reasoning process is based on the concept of intention. In this model, intentions are cases, which have to be retrieved, reused, revised and retained [8], [13]. To achieve both goals, the structure of the CBR system has been designed around the concept of a case.…”

“…In addition, they have the capacity to decide what to do and how to get it according to their attitudes [15], [20], [26]. The paper describes how intelligent Internet systems may be constructed with autonomous agents that reason with the help of a CBR system [8], [14].…”