Abstract
Test case generation is an important process in software testing. However, manual generation of test cases is a time-consuming process. Automation can considerably reduce the time required to create adequate test cases for software testing. Genetic algorithms (GAs) are considered to be effective in this regard. The multiple-searching genetic algorithm (MSGA) uses a modified version of the GA to solve the multicast routing problem in network systems. MSGA can be improved to make it suitable for generating test cases. In this paper, a new algorithm called the enhanced multiple-searching genetic algorithm (EMSGA), which involves a few additional processes for selecting the best chromosomes in the GA process, is proposed. The performance of EMSGA was evaluated through comparison with seven different search-based techniques, including random search. All algorithms were implemented in EvoSuite, which is a tool for automatic generation of test cases. The experimental results showed that EMSGA increased the efficiency of testing when compared with conventional algorithms and could detect more faults. Because of its superior performance compared with that of existing algorithms, EMSGA can enable seamless automation of software testing, thereby facilitating the development of different software packages.
Highlights
Software testing is an important process in the software development life cycle
enhanced multiplesearching genetic algorithm (EMSGA) generated fewer cases, the branch coverage of EMSGA was higher. This means that EMSGA is more efficient than monotonic Genetic algorithms (GAs)
The A12 measure is a comparison of effect size between the EMSGA and the others; if A12 > 0.5, it means EMSGA can beat that algorithm more than 50% of the time
Summary
Software testing is an important process in the software development life cycle. It is performed to investigate the quality of software and to evaluate the risks in software implementation. From previous work [7], MSGA can generate test cases for small to medium scale software but cannot increase the percentage of coverage for complex software. This means test cases generated with MSGA cannot increase the number of executed statements or source code in complex software. While MSGA may be suitable for generating test cases for small to medium scale software, it may not be flexible enough for test case generation for complex software. Some algorithms may be suitable for generating test cases for small to medium scale software but may not succeed in complex cases. We refactored the algorithm to solve the problem of executing the source code for more access to the statements
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
