The Layered Architecture Recovery as a Quadratic Assignment Problem

Belle, Alvine Boaye; El Boussaidi, Ghizlane; Desrosiers, Christian; Kpodjedo, Sègla et Mili, Hafedh (2015). « The Layered Architecture Recovery as a Quadratic Assignment Problem », dans Software Architecture: 9th European Conference, ECSA 2015, Dubrovnik/Cavtat, Croatia, September 7-11, 2015. Proceedings, sous la dir. de Weyns, Danny; Mirandola, Raffaela et Crnkovic, Ivica. Cham, Springer International Publishing, coll. «Lecture Notes in Computer Science », vol. 9278, no 9278, pp. 339-354.

Fichier(s) associé(s) à ce document :
Télécharger (289kB)


Software architecture recovery is a bottom-up process that aims at building high-level views that support the understanding of existing software applications. Many approaches have been proposed to support architecture recovery using various techniques. However, very few approaches are driven by the architectural styles that were used to build the systems under analysis. In this paper, we address the problem of recovering layered views of existing software systems. We re-examine the layered style to extract a set of fundamental principles which encompass a set of constraints that a layered system must conform to at design time and during its evolution. These constraints are used to guide the recovery process of layered architectures. In particular, we translate the problem of recovering the layered architecture into a quadratic assignment problem (QAP) based on these constraints, and we solve the QAP using a heuristic search algorithm. In this paper, we introduce the QAP formulation of the layering recovery and we present and discuss the results of the experimentation with the approach on four open source software systems.

Type: Chapitre de livre
Mots-clés ou Sujets: Software architecture, Architecture recovery, Layered style, Architecture evolution, Quadratic assignment problem.
Unité d'appartenance: Faculté des sciences > Département d'informatique
Déposé par: Hafedh Mili
Date de dépôt: 21 avr. 2016 15:11
Dernière modification: 27 avr. 2016 18:38
Adresse URL :


Voir les statistiques sur cinq ans...