Techreport,
Intelligent Timing Analysis of Real-Time Software
Internal Report, University of Glamorgan School of Computing, (October 1998)
Abstract
The operation speed of Real-Time Software is critical. Timing Analysis is therefore an important part of the design, testing and assessment of Real-Time Software. It measures how well a system matches its specification. New methods such as Genetic Algorithms now make the Dynamic Timing Analysis of systems feasible and are able to replace or supplement standard Static Timing Analysis techniques. These dynamic techniques give much more accurate predictions of a system's run-time behaviour through their analysis of the interactions of the program's input parameters. Experiments with the new methodologies have shown significant improvement in the prediction of timing constraints and confirmed their value for module testing. The applicabilty of the new methodologies to larger and more complex systems will be thoroughly investigated. This will form the basis of the original contribution to knowledge for subsequent work.
