Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
4951857 | Science of Computer Programming | 2016 | 17 Pages |
Abstract
Model-based statistical testing with Markov chain usage model (MCUMs) represents a highly automated test approach. It also boosts test planning by providing relevant parameters that aid decision making. However, MCUMs are not suitable for testing modern systems that allow concurrent streams of use. Concurrent aspects result in a state space explosion in the model. They complicate the model specification and increase the calculation time for parameters. Therefore, we introduce concurrency enhanced usage model (CenUMs). They represent a natural extension to MCUMs by composite states with regions, which are used to formalize concurrent streams of use. This allows a simplified and reasonable specification as it hides the state space explosion. We also provide configurable algorithms for test case generation that take composite states into account. Additionally, a new analysis method is outlined that leads to a significant improvement of the calculation time for parameters. An analysis of its arithmetic complexity and results from experiments are provided.
Keywords
Related Topics
Physical Sciences and Engineering
Computer Science
Computational Theory and Mathematics
Authors
Daniel Homm, Jürgen Eckert, Reinhard German,