Improving the accuracy of COPLIMO to estimate the payoff of a software product line

Software product line engineering pursues the efficient development of families of similar products. COPLIMO is an economic model that relies on COCOMO II to estimate the benefits of adopting a product line approach compared to developing the products one by one. Although COPLIMO is an ideal economic model to support decision making on the incremental development of a product line, it makes some simplifying assumptions that may produce high distortions in the estimates (e.g., COPLIMO takes for granted that all the products have the same size). This paper proposes a COPLIMO reformulation that avoids such assumptions and, consequently, improves the accuracy of the estimates. To support our proposal, we present an algorithm that infers the additional information that our COPLIMO reformulation requires from feature diagrams, which is a widespread notation to model the domain of a product line.