Date: Wed, 20 Nov 1996 19:16:53 GMT Server: Apache/1.0.3 Content-type: text/html Content-length: 1843 Last-modified: Sat, 06 Jul 1996 18:44:27 GMT
This project applies case-based reasoning to the problem of selecting methods for solving sparse linear systems in scientific computation. Because there are no "hard-and-fast" rules to determine which solvers apply, the experience-based reasoning of CBR is a particularly promising method for solver selection. The goal is to develop algorithms that can automatically learn which methods to apply to different problems on different types of hardware, and to publicly distribute an implementation of these algorithms with the solvers being developed as part of the I.U. HyperMatrix project in scientific computation.
Associated Faculty: David B. Leake
Associated Graduate Students: Vikram Subramaniam and David Wilson.
Affiliated Projects: Scientific computation at Indiana University
Support: This research is supported by a systems assistantship from the Indiana University Computer Science Department.
For more information about CBR research at Indiana, click here