Search Based Automatic Test-Data Generation at an Architectural Level

Abstract. The need for effective testing techniques for architectural level descriptions is widely recognised. However, due to the variety of domain-specific architectural description languages, there remains a lack of practical techniques in many application domains. We present a simulation-based testing framework that applies optimisation-based search to achieve high-performance testing for a type of architectural model. The search based automatic test-data generation technique forms the core of the framework. Matlab/Simulink is popularly used in embedded systems engineering as an architectural-level design notation. Our prototype framework is built on Matlab for testing Simulink models. The technology involved should apply to the other architectural notations provided that the notation supports execution or simulation. 1 Automatic Testing at the Architecture Level Software testing is an expensive procedure. It typically consumes more than 50% of the total development budget [1]. Fai...

• Type: Conference
• Year: 2004
• Where: GECCO
• Authors: Yuan Zhan, John A. Clark
• Keywords: Effective Testing Techniques, GECCO 2004, Software Testing, Test-data Generation