Evaluating Variability Modeling Techniques for Supporting Cyber-Physical System Product Line Engineering

Safdar Aqeel Safdar1, Tao Yue1,2, Shaukat Ali1, Hong Lu1

1Simula Research Laboratory, Oslo, Norway; 2Department of Informatics, University of Oslo, Oslo, Norway
Contact: {safdar, tao, shaukat, honglu}@simula.no

Abstract

Modern society is increasingly dependent on Cyber-Physical Systems (CPSs) in diverse domains such as aerospace, energy and healthcare. Employing Product Line Engineering (PLE) in CPSs is cost-effective in terms of reducing production cost, and achieving high productivity of a CPS development process as well as higher quality of produced CPSs. To apply CPS PLE in practice, one needs to first select an appropriate variability modeling technique (VMT), with which variabilities of a CPS Product Line (PL) can be specified. In this paper, we proposed a set of basic and CPS-specific variation point (VP) types and modeling requirements for proposing CPS-specific VMTs. Based on the proposed set of VP types (basic and CPS-specific) and modeling requirements, we evaluated four VMTs: Feature Modeling, Cardinality Based Feature Modeling, Common Variability Language, and SimPL (a variability modeling technique dedicated to CPS PLE), with a real-world case study. Evaluation results show that none of the selected VMTs can capture all the basic and CPS-specific VP and meet all the modeling requirements. Therefore, there is a need to extend the existing techniques or propose new ones to satisfy all the requirements.

Case Study Models:

As mentioned in the paper [1], we model the Material Handeling (MHS) System using selected four variability modeling techniques including Feature Model (FM), Cardinality Based Feature Model (CBFM), SimPL, and Common Variability Language (CVL).

Data related with the experiment

Download FM The MHS models for Feature Model. To model the case study with Feature Model we used Pure::Variant tool [2].

Download CBFM The MHS models for Cardinality BasedFeature Model. To model the case study with Cardinality BasedFeature Model we used fmP tool [3].

Download SimPL The MHS models for SimPL. To model the case study with SimPL we used IBM Rational Software Architect tool [4].

Download CVL The MHS models for Common Variability Language. To model the case study with CVL we used CVL tool [5].