Computer-aided software evolution based on inferred dependencies

Abstract

The major problem addressed by this research is how to automate parts of software evolution using dependency rules, especially for large and complex real time embedded systems. The main topics of this study are the development of a Relational Hypergraph model (RH model) and the design of a Computer Aided Software Evolution System (CASES). The goals of this dissertation are to explore the existing issues, to formalize software evolution, to reuse software evolution components, and to build a dependency computing model. We have resolved parts of essential software evolution issues in the following categories: software evolution process, software evolution traceability, software evolution description, software evolution management, and software evolution control. The RH model can realize automated software evolution in multi-dimensional phases, such as software prototype or product demo, issue analysis, requirement analysis, specification design, module implementation, program integration, and software product implementation. Many types of software evolution objects in each phase, and dependencies among these objects have been defined to describe software evolution processes.