A Graph Based Feature Selection Approach for Configuring Software Product Lines

Project: Research

Project Details

Description

A Software Product line (SPL) is configuration centric with a focus on developing a collection of related software products, all of which share some core functionality, and differ in some specific features. SPL use a feature model to specify the commonalities and variabilities in terms of software features to identify and develop reusable software assets. Feature selection is a key process to derive new SPL configurations that aims to either minimize or maximize an object function subject to a set of constraints. To date, feature selection techniques have focused on finding an optimal solution to objective functions such as cost, subject to resource constraints. However, the existing approaches have not considered the structural relationships and configuration dependencies encoded in a feature model, leaving open the question of how best to optimize SPL feature selection in presence of feature interdependencies. In this research project, we aim to develop a feature selection technique that consolidates interdepend features into related clusters and uses a genetic search algorithm to find a near optimal solution for feature selection subject to feature priority and product integrity constraints. This project is expected to provide a solution to SPL feature selection problem such that it helps a developer to analyse interdependencies and select suitable features for SPL configurations. We also plan to present application of our approach to real world case studies.
StatusFinished
Effective start/end date11/04/1611/10/17

Fingerprint

Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.