Case-Based Travel Agent

Abstract

With the explosion of the Internet, we are getting more and more information and it becomes difficult to digest. Agents exist to help the users. An agent is a component of software (or hardware) that is capable of accomplishing tasks on behalf of its user. To accomplish these tasks, it should be able to adapt itself to the user and should also be able to learn from his decisions. Most of them use a rule-based system and it appears that they are not really adapted to the user. A case-based reasoning system seems to be a better approach. Rule-based systems need a good definition of the system and are not able to learn very well from the user choices. What is more, it is not able to handle missing information or unexpected values. On the other hand, an agent build with a case-based reasoning system should be able to solve these problems. In order to show the benefits of a case-based reasoning system for an agent, one of the FIPA (Foundation for Intelligent Physical Agents) scenarios will be used as a context for this dissertation: the Travel Assistance Scenario. This scenario involves different kinds of agents that interact in order to plan a journey for a user. In this dissertation, a case-based system will be implemented using a Case-Retrieval Net (CRN) for retrieval. It will then be used to demonstrate the flexibility and adaptability of this kind of system.