Self-Adaptation Driven by SysML and Goal Models -- A Literature Review

Amal Ahmed Anda,Daniel Amyot

doi:10.37190/e-inf220101

Abstract

Background: Socio-cyber-physical systems (SCPSs) are a type of cyber-physical systems with social concerns. Many SCPSs, such as smart homes, must be able to adapt to reach an optimal symbiosis with users and their contexts. The Systems Modeling Language (SysML) is frequently used to specify ordinary CPSs, whereas goal modeling is a requirements engineering approach used to describe and reason about social concerns. Objective: This paper aims to assess existing modeling techniques that support adaptation in SCPSs, and in particular those that integrate SysML with goal modeling. Method: A systematic literature review presents the main contributions of 52 English articles selected from five databases that use both SysML and goal models (17 techniques), SysML models only (11 techniques), or goal models only (8 techniques) for analysis and self-adaptation. Result: Existing techniques have provided increasingly better modeling support for adaptation in a SCPS context, but overall analysis support remains weak. The techniques that combine SysML and goal modeling offer interesting benefits by tracing goals to SysML (requirements) diagrams and influencing the generation of predefined adaptation strategies for expected contexts, but few target adaptation explicitly and most still suffer from a partial coverage of important goal modeling concepts and of traceability management issues.

Highlights

Cyber-physical systems (CPSs) are systems that tightly “integrate physical, software, and network aspects in a sometimes adverse physical environment” [1]
This paper reviewed 52 publications and assessed methods that integrate goal models with Systems Modeling Language (SysML) models to support runtime self-adaption, with a consideration for the Socio-cyber-physical systems (SCPSs) context
The review answers many questions of broad interest both to researchers and to practitioners who are considering the use of goal models, SysML models, PREVIEW or both in SCPSs or self-adaptive systems contexts

Summary

Introduction

Cyber-physical systems (CPSs) are systems that tightly “integrate physical, software, and network aspects in a sometimes adverse physical environment” [1]. They are composed of hybrid components such as hardware (e.g., sensors, devices, and networks) and software, which can even be integrated at runtime. Many SCPSs should ideally be able to adapt to changing conditions in order to reach an optimal symbiosis with users (and other stakeholders) and their contexts [2]. Objective: This paper aims to assess existing modeling techniques that support adaptation in SCPSs, and in particular those that integrate SysML with goal modeling. The techniques that combine SysML and goal modeling offer interesting benefits by tracing goals to SysML (requirements) diagrams and influencing the generation of predefined adaptation strategies for expected contexts, but few target adaptation explicitly and most still suffer from a partial coverage of important goal modeling concepts and of traceability management issues

Objectives

Methods

Conclusion