A similarity-based view to distributed information retrieval with mobile agents

Although advanced and innovative technologies have been used to retrieve information, the need of advanced, new techniques for Web information retrieval is extant. One of the main difficulties is to return accurate information that fully describes the human request. Standard Web search engines provide two mechanisms to handle a user's query: to input the search keywords or to restrict the searching within a certain topic. These approaches appear too weak to cope with such a vast amount of information and with such a rich human expressiveness. Our approach to Web information retrieval consists in merging different available technologies and paradigms: mobile computation is exploited as effective approach to obtain updated results by dispatching the agents directly on Web resources. Agents are equipped with deductive behavior so to be autonomous in applying inference-based reasoning. A similarity model is embedded into the reasoning engine in order to correctly process the approximation level established at user level side.