Towards automated software architectures design using model transformations and evolutionary algorithms
Rui Li, Michel R. V. Chaudron, René C. Ladan · 2010
The design of software architecture is one of the large challenges in modern software engineering. It requires software architects to address a large number of non-functional requirements related to performance, safety, availability and cost quality attributes. Moreover, these quality attributes often conflict with each other, for instance, improving system performance often needs more powerful hardware nodes, which increases the production cost and power consumption in the meantime. In this paper, we present the PETUT-MOO tool (Performance-Enhancing Tool using UML Transformations and Multi-objective Optimizations) which can analyze a given software architecture, propose alternatives to it, and do architecture optimization to improve its non-functional properties in an automatic way.