Evaluation of program test methods using mutation analysis

Abstract

The information about the effectiveness of program test methods aids to increase one``s confidence in the programs tested by them. This thesis presents results of an experimental evaluation of branch testing, weak mutation testing, random testing, and specification testing. In addition the effect of special/extremal values testing is analyzed also. Mutation analysis provides a quantitative measure of comparison for the quality of a set of test cases. Test data sets are generated according to the requirements of each of the strategies and executed by a mutation system on sample programs, then the execution results are analyzed.