By Topic

Efficient evaluation of multifactor dependent system performance using fractional factorial design

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$33 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

2 Author(s)
T. Berling ; Ericsson Microwave Syst. AB, Molndal, Sweden ; P. Runeson

Performance of computer-based systems may depend on many different factors, internal and external. In order to design a system to have the desired performance or to validate that the system has the required performance, the effect of the influencing factors must be known. Common methods give no or little guidance on how to vary the factors during prototyping or validation. Varying the factors in all possible combinations would be too expensive and too time-consuming. This paper introduces a systematic approach to the prototyping and the validation of a system's performance, by treating the prototyping or validation as an experiment, in which the fractional factorial design methodology is commonly used. To show that this is possible, a case study evaluating the influencing factors of the false and real target rate of a radar system is described. Our findings show that prototyping and validation of system performance become structured and effective when using the fractional factorial design. The methodology enables planning, performance, structured analysis, and gives guidance for appropriate test cases. The methodology yields not only main factors, but also interacting factors. The effort is minimized for finding the results, due to the methodology. The case study shows that after 112 test cases, of 1024 possible, the knowledge gained was enough to draw conclusions on the effects and interactions of 10 factors. This is a reduction with a factor 5-9 compared to alternative methods.

Published in:

IEEE Transactions on Software Engineering  (Volume:29 ,  Issue: 9 )