Efficient information retrieval on the World Wide Web using adaptable and mobile Java agents

When searching for a specific item of information across a vast number of databases accessible on the World Wide Web (WWW), one is faced with the problem of obtaining an intelligent response to a topic query in a feasible period of time. One such complicated grouping of information is the United States legal code where the typical query can bring back tens of thousands of related documents spanning numerous law repositories. To be practical, efficient information retrieval is critical, i.e., decisions must be made as to which databases to search and which of the information hits to return. One approach is to search the repositories in parallel using intelligent and adaptive mobile software agents. In this paper, we describe various tradeoffs and the feasibility of using adaptable mobile Java agents to solve the problem of efficiently searching and filtering through large amounts of information on the WWW. We show that for a legal code keyword query, mobile Java agents can be made to adapt (e.g., refocus their search criteria and reroute their search path) to efficiently retrieve the desired information within a new optimized search region of the WWW.