An Evolutionary Approach for Optimisation of State-Based Test Suites for Software Systems

Document Type

Conference Proceeding


International Association for Computer and Information Science


Faculty of Computing, Health and Science


School of Computer and Security Science




Li, H., & Lam, C. P. (2003). An evolutionary approach for optimisation of state-based test suites for software systems. Proceedings of the ACIS Fourth International Conference on Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing (SNPD'03). (pp. 226-233). Mount Pleasant, Michigan. International Association for Computer and Information Science. Available here


There are many different approaches to generate the test data for software testing, one of the frequently used approaches is the state-based test data generation. In general, test coverage specified by the per-determined testing requirements is frequently used to ascertain the quality of software testing. Test Suites are automatically generated accordingly to cover the required coverage. Properly generated test suite is one of the key assurances to the software system quality. However, automatic test data generation is always one of the most problematic issues in state-based testing. More test cases are included in a test suite, it is more likely for the test suite to satisfy the required testing coverage. However, more test cases may inevitably introduce higher testing cost due to the inclusion of many redundant test cases. It is desired that the cost associated with redundant test cases can be reduced without sacrificing the quality of the software. This paper proposes an evolutionary approach to optimise the state-based test suites. The optimisation is achieved by reducing the number of redundant test cases while still guaranteeing the required transition coverage of the test suite.