A component-based system is a set of entities that work together in well-defined ways to satisfy a given requirement specified by the stakeholders for the system. This requirement can be modeled as a set of combinations of traits, which represent acceptable alternatives for providing the required functionality. A system satisfies its requirement if and only if it provides one of the required sets of traits in its entirety. Beyond the requirement, system stakeholders may also have preferences with respect to optional functionality that could be provided by a system, tradeoffs between non-functional properties, or other system design options. This work focuses on integrating support for both qualitative preference reasoning and formal verification into the component-based system design process in order to choose a set of components for the system that, when composed, will (1) satisfy the stakeholders\u27 requirement for the system and (2) provide a set of traits that is optimal with respect to the given preferences. Our primary research objective is to develop a generic, modular, end-to-end framework for developing component-based systems of any type which are correct according to the system requirement and most preferred with respect to the stakeholders\u27 preferences. Applications of the framework to problems in Web service composition, goal-oriented requirements engineering, and other areas will be discussed, along with future work toward integrating multi-stakeholder preference reasoning and partial satisfaction of traits into the framework.
展开▼
