Abstract. Evolutionary Testing (ET) has been shown to be very successful for testing real world applications [10]. The original ET approach focuses on searching for a high coverage of the test object by generating separate inputs for single function calls. We have identified a large set of real world application for which this approach does not perform well because only sequential calls of the tested function can reach a high structural coverage (white box test) or can check functional behavior (black box tests). Especially, control software which is responsible for controlling and constraining a system cannot be tested successfully with ET. Such software is characterized by storing internal data during a sequence of calls. In this paper we present the Evolutionary Sequence Testing approach for white box and black box tests. For automatic sequence testing, a fitness function for the application of ET will be introduced, which allows the optimization of input sequences that reach a high coverage of the software under test. The authors also present a new compact description for the generation of real-world input sequences for functional testing. A set of objective functions to evaluate the test output of systems under test have been developed. These approaches are currently used for the structural and safety testing of car control systems.
A hybrid specification language µSZ, in which the dynamic behaviour of a system is described using Statecharts and the data and the data transformations are described using Z, has been developed for the specification of embedded systems. This paper describes an approach to testing from a deterministic sequential specification written in µSZ. By considering the Z specifications of the operations, the extended finite state machine (EFSM) defined by the Statechart can be rewritten to produce an EFSM that has a number of properties that simplify test generation. Test generation algorithms are introduced and applied to an example. While this paper considers µSZ specifications, the approaches described might be applied whenever the specification is an EFSM whose states and transitions are specified using a language similar to Z.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.