SOFTWARE RELIABILITY FORECASTING FOR ADAPTIVE FAULT TOLERANCE ALGORITHMS
Abstract: The seminar presents adaptive methods based on statistical
measurements in order to choose dinamically an appropriate algorithm for
fault tolerant systems. Statistical methods are necessary for the prediction
of the future failure behavior and operational environment. The originality
of our approach is to adapt automatically the fault management according
to the reliability forecasting. A software program SOFTRELY-WEB_1.1 was
implemented on Java to illustrate the software reliability prediction.
This seminar was held at the Department of Computer Science, Royal Holloway,
University of London on 20 May 1998