Regression Test Suites Optimization Using TLBO Based Adaptive Neuro-Fuzzy Inference System

Abstract

Regression testing is among the major activities in Software Engineering that is done whenever modifications are made in a software. New test cases are required to be added in current test suite for checking the enhanced functionalities. But, the size of test suite increase as new test cases are added and it becomes un-efficient because of the occurrence of redundant, broken and obsolete test cases. For that reason, it results in additional time and budget to run all these test cases. Therefore, in order to overcome the problem of time as well as budget constraint, it is required to optimize the entire test suite. Many researchers have proposed computational intelligence and conventional based approaches for dealing with this problem and they have achieved optimized test suite by selecting, minimizing or reducing, and prioritizing test cases. Currently, most of the approaches dealing with optimization are static in nature and they do not dynamically modify the test cases. But, it is mandatory to use dynamic approaches for optimization due to the advancements in information technology and associated market challenges. Therefore, we have proposed an Adaptive Neuro Fuzzy Inference System (ANFIS) that is tuned with Teaching Learning Based Optimization (TLBO) algorithm, for optimizing the regression test suites. NeuroFuzzy Modeling (NFM) is a dynamic approach that is used for describing the system through ifthen else rules and network structure is utilized for its representation. For dealing with uncertain values of input, these neuro-fuzzy based models provide effective methods along with improved consistency. They also exhibit good property of generalization and their interpretation is done by experts. In this dissertation, two benchmark case studies have been used and controlled experimentation have been performed for optimization of test cases. The validation and comparison of our approach has been done with GA-ANFIS, PSO-ANFIS, FA-ANFIS and HSANFIS. From our results, it has been concluded that proposed TLBO-ANFIS performs better than all of these approaches