Multiple software product lines to configure applications of internet of things. Issue 2 (1st April 2020)

• Record Type: Journal Article
• Authors: Trujillo-Tzanahua, Guadalupe-Isaura; Juárez-Martínez, Ulises; Aguilar-Lasserre, Alberto-Alfonso; Cortés-Verdín, María-Karen; Azzaro-Pantel, Catherine
• Is Part Of: IET software. Volume 14:Issue 2(2020)
• Page Start: 165
• Page End: 175
• Publication Date: 2020-04-01
• Subjects: Internet of Things -- genetic algorithms -- software engineering -- mass production -- production engineering computing -- software reusability -- software product lines -- search problems -- cost reduction; multiple software product lines -- industrial production lines -- reuse schemes -- mass production -- universal feature model -- multiobjective optimisation model -- search-based software engineering technique -- Internet of Things -- development cost minimisation -- genetic algorithms; Computer software -- Periodicals; Software engineering -- Periodicals; 005.1
• DOI: 10.1049/iet-sen.2019.0032
• Languages: English
• ISSNs: 1751-8806

Abstract

Abstract : Software product lines (SPL) emulate the industrial production lines that are capable of generating large volumes of products through reuse schemes and mass production. A multi product line (MPL) aims to reuse of several SPL. Feature models are often used to manage the existing resources of SPLs and define valid products through notations and relationships such as mandatory, optional, and alternative features. The main contribution of this study is a method to manage the variability of multiple SPL and generate a new portfolio of products for Internet of Things (IoT). For this, the problem of developing a universal feature model (FM) for an MPL from merging the FMs of the individual SPLs with a Search‐Based Software Engineering (SBSE) technique is addressed. In addition, the authors propose a multi‐objective optimisation model to maximise the reusability and compatibility between features and minimise the development cost. The model facilitates the design of an MPL‐feature model. Authors' empirical results show that the proposed model solved by genetic algorithms allows to configure a variety of software products and to determine the scope of the MPL.