Developing a conceptual model and simulator for dependable software engineering

Software development projects necessitate many decision-makings by project managers. Focusing on aspects like software development time and effort the reviewing of all possible influence factors becomes more and more complex. If project managers consider additionally dependability, a decision-making based on empirical data is to be advised against. One of the problems is that many factors are unpredictable as they show changes over time. This thesis tackles the problem of decision-making by combining empirical data with software process simulation. Different Analytic Quality Engineering Techniques (inspection, architecture evaluation and testing) will offer project managers' strategies how an increase of dependability can be achieved within single system engineering or product line engineering. Associated aspects like benefit interrelated to effort and time will be included to allow a faithful context-representation. The simulation model is developed based on a conceptual model, which represents the functionality and interrelations of the simulation model. Empirical data of former projects [RUS98] [LIN+05] and assumptions were consulted to establish the model. An exemplified software project which takes two alternative quality engineering strategies as input, will illustrate the functionality of the simulation model. Using the model, a manager can compare both strategies and choose the one fitting best for his subsequent project. Conclusions regarding the effectiveness of the simulation model will complete this work.