R. Zhao and C. Li (PRC)
Test data generation, genetic algorithms, fitness function, and non-numerical data
Non-numerical data is an important element in programming. A problem that needs further research is how to automatically generate software test data for non-numerical variables. This paper discusses a genetic algorithm that can automatically generate test cases of non-numerical data to check a given program path, and defines a path fitness function and a character fitness function with respect to two different types of character string predicate. The algorithm takes a given path as a target path, and then automatically selects a path fitness function or a character fitness function according to the predicates on target path to conduct test data generation. The effectiveness of this approach is examined on a number of programs, and the preliminary experimental results show that this approach is feasible and effective.
Important Links:
Go Back
