A Critical Review on Automated Test Case Generation for Conducting Combinatorial Testing Using Particle Swarm Optimization

Abstract

In software development life cycle, testing plays the significant role to verify requirement specification, analysis, design, coding and to estimate the reliability of software system. A test manager can write a set of test cases manually for the smaller software systems. However, for the extensive software system, normally the size of test suite is large, and the test suite is prone to an error committed like omissions of important test cases, duplication of some test cases and contradicting test cases etc. When test cases are generated automatically by a tool in an intelligent way, test errors can be eliminated. In addition, it is even possible to reduce the size of test suite and thereby to decrease the cost & time of software testing. It is a challenging job to reduce test suite size. When there are interacting inputs of Software under Test (SUT), combinatorial testing is highly essential to ensure higher reliability from 72 % to 91 % or even more than that. A meta-heuristic algorithm like Particle Swarm Optimization (PSO) solves optimization problem of automated combinatorial test case generation. Many authors have contributed in the field of combinatorial test case generation using PSO algorithms. We have reviewed some important research papers on automated test case generation for combinatorial testing using PSO. This paper provides a critical review of use of PSO and its variants for solving the classical optimization problem of automatic test case generation for conducting combinatorial testing.

Combinatorial Testing Approaches: A Systematic Review

“…After examining the requirements, the parameters and values were used to describe the input to the system. A set of tests was conducted to cover all possible combinations of the specified parameters [9]. Following the analysis requirements, the parameters and values describe the system input, and every possible combination is tested.…”

“…The investigation closes with a discussion on the matter and prospects for constrained testing. Prakash et al [9] offered a step-by-step testing function to generate reviews of a novel algorithm. A significant amount of effort has gone into understanding that number in the past, meaning that more work can be done in the future.…”

“…This means that articles that do not focus on combinatorial interactions are overlooked by combinatorial testing scientists. Various studies have been conducted on these documents [8,9]. This was also discussed because of the significance of combinatorial interaction testing for the implementation of the constraint control flow.…”