Sip

Abstract

A feature model specifies the sets of features that define valid products in a software product line. Recent work has considered the problem of choosing optimal products from a feature model based on a set of user preferences, with this being represented as a many-objective optimisation problem. This problem has been found to be difficult for a purely search-based approach, leading to classical many-objective optimisation algorithms being enhanced by either adding in a valid product as a seed or by introducing additional mutation and replacement operators that use a SAT solver. In this paper we instead enhance the search in two ways: by providing a novel representation and also by optimising first on the number of constraints that hold and only then on the other objectives. In the evaluation we also used feature models with realistic attributes, in contrast to previous work that used randomly generated attribute values. The results of experiments were promising, with the proposed (SIP) method returning valid products with six published feature models and a randomly generated feature model with 10,000 features. For the model with 10,000 features the search took only a few minutes. CCS Concepts: r Software and its engineering → Software product lines; r Mathematics of computing → Optimization with randomized search heuristics;

A novel aggregation-based dominance for Pareto-based evolutionary algorithms to configure software product lines

“…3) Our experiments show that compared to the original SIP framework-based algorithms used in [16], the performance of both NSGA-II-ADO and SPEA2+SDE-ADO were improved as measured by hypervolume for 7/9 and 8/9 SPLs, respectively. Furthermore, SPEA2+SDE-ADO outperforms the other 4 state-of-the-art MOEAs in terms of hypervolume metric for 8 out of 9 SPLs.…”

“…Evolution computation techniques, such as particle swarm optimization (PSO) and evolutionary algorithms (EA) have been successfully applied in many real-world optimization problems due to their population-based metaheuristic that allows to search for a set of optimal solutions in a single run [7][8][9][10][11][12]. In the past decade, there have been many studies that adopt different multi-objective evolutionary algorithms (MOEAs) as automatic configuration approaches to solve the optimal feature selection problem [6,[13][14][15][16][17][18][19].…”

“…Henard et al [14] introduced two smart evolutionary operators into IBEA, called SATIBEA where an invalid solution (a software product configuration with constraint violations) is identified and fixed by using the SAT solver. Hierons et al [16] presented a ShrInk Prioritize (SIP) framwork for different type of MOEAs. In particular, the novel encoding method is used to identify prunable features and the (1 + n) approach is used to first optimize the number of constraint violations and then other objectives at the same time to guide the search to feasible space.…”

Constraint programming heuristics for configuring optimal products in multi product lines

“…We consider a set of FMs found in the state-of-the-art. Despite the absence of MPLs in the literature, this experiment is done in order to test our results against case studies that are usually considered in the experimentation phase of other performance-concerned approaches [11,12,28,[37][38][39] .…”

A systematic literature review on the semi-automatic configuration of extended product lines

“…An analysis that considers both sets is unnecessary, since core features are present in all products, and dead features are never included. Hierons et al (2016) also remove selected parent features if one or more child features are selected. These features are added back when the remaining decisions are defined.…”