International Journal of Computer Theory and Engineering

Abstract

Software system should be reliable and available failing which huge losses may incur. To achieve these objectives a thorough testing is required. Adequacy of test cases is the key to the success. Despite the availability of a number of adequacy criteria, deterministic approaches to testing are not sufficient consequential to the need of automatic random and anti-random testing. Our research uses a novel method for the development of n-version of the software by creating the different mutation in software and test cases generation using the Genetic Algorithm. Its purpose is to eliminate software faults as possible by using lesser test cases in the testing phase. The test case generated by the use of Genetic Algorithm (GA) is compared with the results of totally random generated test cases. The method was applied to the specification of a sorting of array problem. The advantage of this analysis is that when we produce multiple versions, reliability of the software is likely to be better than if a single version is developed. The N-version software testing will helps to reduce the possibility of mistakes and inconsistencies in the process of software development and testing and the number of test cases required during the testing phase of the software system. In this paper a technique of generating the test cases and doing the testing automatically, employing genetic algorithm and Back-to-Back testing has been discussed.