Mobile Information Agents
Abstract
The main goals for this project were to develop automated information organization algorithms, and to integrate the information organization algorithms in a mobile agent platform. The main objective was to investigate and demonstrate the value of a paradigm of computation in heterogeneous distributed systems with non-permanent network connections, in which mobile agents bring the computation to the data. As a result, a system called D'Agents that supports mobile agents has been developed. D'Agents is especially suited to distributed information access experiments in a network of mobile computers, such as laptops, palmtops, and other wireless devices. In addition, this effort has developed, implemented, and evaluated an information organization algorithm called the star algorithm. The star algorithm gives an organization of collection into clusters. Results for this effort include: 1. Information overload is a serious problem and efficient automatic information organization algorithms are useful in addressing this problem. 2. The Star Clustering algorithms: a. is the best performing algorithm for large-scale information organization, b. can be used in an on-line or off-line fashion and has several scalable extensions, c. has been analyzed and this effort's large-scale experiments math the theory, d. can be used for filtering applications and for persistent queries. 3. By combining the Star clustering algorithm with the power of mobile agent system, we increase system performance dramatically.
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 01, 2001
- Accession Number
- ADA390712
Entities
People
- Daniela L. Rus
Organizations
- Dartmouth College