SOFTWARE TEST DATA GENERATION USING GENETIC ALGORITHM

Abstract

A small change in the software system may lead to malfunctioning of the existing software system. Thus, there arises the need for efficient and selective Software Testing. Software Testing is the process of testing a software system after it has undergone development and up -gradation. It aims to detect faults, if any, that may have been introduced into the software system as a result of these changes. In software engineering research, Genetic Algorithm is used to generate and refine the automated test-data for developed software product in an efficient & quick manner. Genetic Algorithm is an adaptive search technique that improves the software testing process in an efficient manner. It improves the testing-automation, where traditional methods are considered too complex and time consuming. In this thesis we propose and validate a test case refinement framework based on Genetic Algorithm (GA) Further, comparison of test data generation and refinement using Genetic Algorithm as well as Random Algorithm is discussed and presented. A tool has been developed which gives comparison table as well as comparison graph of fitness values. The proposed and developed tool can be used with different types of software systems. Based on the results; it is concluded that the Genetic Algorithm proves to be better than Random Algorithm in terms of automated generation and refinement of software test data.