Abstract

Real-time reactive systems are largely event-driven, interact intensively and continuously with the environment through stimulus-response behavior, and are regulated by strict timing constraints. Examples of such systems include alarm systems, air traffic control systems, nuclear reactor control systems and telecommunication systems; applications involving real-time reactive software play a mission-critical role in the defense industry. Real-time reactive systems are inherently complex. The complexity pervades through the different phases of software development, deployment, and maintenance. Applying formal methods in the development process is an effective way for dealing with the complexity, and for quality assurance. One of the goals is to assess the quality of such systems starting from the earlier phases of their life cycle. The integration of the quality measurement into the development framework provides feedback to the system developers in order to effectively control the development processes and to obtain high reliability of a final product. Thus, quality control is a must when safety-critical real-time reactive systems are developed. The quality assessment must be regarded as a support for controlling the process of software development in order to guarantee the final quality. The aim of the thesis is to correctly apply the measurement theory to formal description of real-time software upon which we can base models of object-oriented software measurement. In order to create the framework for the present work, we are surveying the theoretical approaches to software measurement. The novelties of the quality measurement methodology are in the theoretical basis and a practical automated measurement data generation process for real-time reactive systems. The proposed approach is applicable to real-time reactive systems modeled as timed labeled transition systems.