Basis Path Testing Using SGA & HGA with ExLB Fitness Function

Abstract

Software testing is the most significant analytic quality assurance for software products, but it is very expensive and time consuming process. This limitation is overcome by automatic testing to reduce high cost and to increase reliability & efficiency as compared to manual testing. Basis path testing is a coverage criterion of software testing that can detect almost sixty five percent of errors in program under test. In this paper a new fitness function has been proposed named as Extended Level Branch (ExLB) Fitness function for basis path testing using simple genetic algorithm (SGA) and hybrid genetic algorithm (hill climbing with selection operator). Using a triangle classifier as program under test, performance of SGA with Simply Combined Fitness Function, SGA with ExLB Fitness Function and HGA with ExLB Fitness Function have been compared using MATLAB. Experimental results showed that SGA with ExLB Fitness Function (proposed approach) performs better than the SGA with Simply Combined Fitness Function and HGA using ExLB Fitness Function is better than all these approaches in terms of test data generation under basis path coverage criteria.