Optimal Component Selection for Component-Based Systems

In Component-based Software (CBS) development, it is desirable to choose software components that provide all necessary functionalities and at the same time optimize certain nonfunctional attributes of the system (for example, system cost). In this paper we investigate the problem of selecting software components to optimize one or more nonfunctional attributes of a CBS. We approach the problem through the lexicographic multi-objective optimization perspective and develop a scheme that produces Pareto-optimal solutions. Furthermore we show that the Component Selection Problem (CSP) can be solved in polynomial time if the components are connected by serial interfaces and all the objectives are to be minimized, whereas the corresponding maximization problem is NP-hard.