Cramer-von Mises Tests for the Compatibility of Two Software Operating Environments

Summary: [This abstract is based on the authors' abstract.] More precise inferences about impending software failures can be obtained when the same reliability model that fits failure data from the test interval also fits data from the field interval However, if the field and test environments differ significantly in terms of how the software is used, a single model for the pooled data may not be adequate. A hypothesis based on a Cramer-von Mises (CvM) test procedure is developed within the context of a nonhomogeneous Poisson process software reliability model. The CvM test has an advantage over a previously-proposed likelihood ratio test in that it does not require that the class of alternatives be specified.