[Updated on Apr. 18] Integration of CiNii Articles into CiNii Research


IPSJ Open Access

Bibliographic Information

Other Title
  • A Strategy of Test Case Prioritization for Software Testing Using Probability Density Function

Search this article



Software testing is a method of validating that a software system operates as intended by actually running it. In software testing, it is important how quickly to find defects, and therefore one of the areas under study is test case prioritization (TCP). Coverage-based TCP prioritizes by using coverage information that indicates which part of the code each test covers. However in existing prioritization strategy the optimal strategy and parameters depend on targets. In this paper, we propose a new strategy in TCP, which gives priority to tests with high anticipated value of the number of detection of newly found defects. Assuming the expectation value of the defect can be calculated using the probability density function representing the distribution of the probability that a certain defect is detected when executing a certain unit, we have estimated this probability density function from the coverage information and calculate the expected value of the defect under the assumption that its distribution follows the beta distribution. For the proposed strategy and existing strategy, we have conducted experiments on 40 C programs and 28 Java programs, and compared them. As a result of the experiment, TCP performance is improved in significant difference by t-test by the proposed strategy when using statement coverage.


Related Projects

See more


Report a problem

Back to top