CCIC: Clustering analysis classes to identify software components

Component identification during software design phase denotes a process of partitioning the functionalities of a system into distinct components. Several component identification methods have been proposed that cannot be customized to software architect’s preferences. In this paper, we propose a clu...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Information and software technology Jg. 57; S. 329 - 351
Hauptverfasser: Hasheminejad, S.M.H., Jalili, S.
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Amsterdam Elsevier B.V 01.01.2015
Elsevier Science Ltd
Schlagworte:
Architecture
Classification
Cluster analysis
Clustering
Component identification
Components
Computer programs
Custom design
Customizing
Design engineering
Hierarchical evolutionary algorithm
Identification methods
Software
Software architect’s preferences
Software industry
Studies
Software architect’s preferences
Hierarchical evolutionary algorithm
Component identification
ISSN:0950-5849, 1873-6025
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Component identification during software design phase denotes a process of partitioning the functionalities of a system into distinct components. Several component identification methods have been proposed that cannot be customized to software architect’s preferences. In this paper, we propose a clustering-based method by the name of CCIC (Clustering analysis Classes to Identify software Components) to identify logical components from analysis classes according to software architect’s preferences. CCIC uses a customized HEA (Hierarchical Evolutionary Algorithm) to automatically classify analysis classes into appropriate logical components and avoid the problem of searching for the proper number of components. Furthermore, it allows software architects to determine the constraints in their deployment and implementation framework. A series of experiments were conducted for four real-world case studies according to various proposed weighting schemes. According to experimental results, it is concluded that CCIC can identify more cohesive and independent components with respect to software architect’s preferences in comparison with the existing component identification methods such as FCA-based and CRUD-based methods.
Bibliographie:SourceType-Scholarly Journals-1
ObjectType-Feature-1
content type line 14
ObjectType-Article-1
ObjectType-Feature-2
content type line 23
ISSN:0950-5849
1873-6025
DOI:10.1016/j.infsof.2014.05.013