This document summarizes a research paper that proposes enhancements to current model-driven engineering methodologies for vehicle software development. Specifically, it proposes using uncertainty modeling to express unknown design decisions in a model of the software architecture. This would allow timing analysis and other evaluations to be performed on the model before finalizing design choices. It suggests that modeling uncertainty could improve analysis performance and allow exploring design tradeoffs between factors like timing and energy consumption. The goal is to support design-space exploration earlier in the development process using model-driven techniques.

1. Towards Design-space Exploration of
Component Chains in
Vehicle Software
Alessio Bucaioni, Antonio Cicchetti, Federico Ciccozzi, Saad
Mubeen, Alfonso Pierantonio and Mikael Sjödin
01-09-2016
SEAA WiP 2016
Arcticus Systems

2. 2
INTRODUCTION
Past Present

3. 3
Bran SelicFather of Real-Time UML
“As our systems grow in complexity
traditional code-centric development
methods are becoming intractable”
INTRODUCTION

4. 4
Model-driven engineering :
abstraction + automation
INTRODUCTION

5. 5
We defined a methodology which seamlessly links
EAST-ADL design and implementation levels to
enable end-to-end delay analysis at design level
BACKGROUND
Alessio Bucaioni, Antonio Cicchetti, Federico Ciccozzi, Romina Eramo ,
Saad Mubeen, Mikael Sjödin
Anticipating Implementation-Level Timing Analysis for Driving Design-Level Decisions in
EAST-ADL
In Proceedings of the International Workshop on Modelling in Automotive Software
Engineering co-located with ACM/IEEE 18th International Conference on Model Driven
Engineering Languages and Systems

6. 6
BACKGROUND
Uncertaint
y

7. 7
BACKGROUND
Uncertainty is due to gaps in domain knowledge,
disagreements between stakeholders, or unresolved
decisions

8. 8
Expressing uncertainty permits a modeler to
specify domain information without prematurely
committing to information she is still uncertain
about, until later refinements can add it.
BACKGROUND

9. BACKGROUND – CURRENT METHODOLOGY
9
Analysis
results
u-JTL
Timing analysis
& filter
Analysis
results
In-place model
transformation
DesginlevelImplementationlevel
EAST-ADL
design model
EAST-ADL
design MM
C2
u-Rubus
model
RubusMM
C2
C2
u-Rubus model annotated with
analysis results
u-JTL
Rubus
model
Rubus
models
u-RubusMM
C2

10. PROBLEM FORMULATION
10
Although notations for dealing with uncertainty
already exist, they are still not fully integrated with
other model-based techniques.
A smooth interplay of these techniques is essential for fully
enabling their benefits and boosting their adoption.

11. PROPOSED ENHANCEMENTS
11
Analysis
results
u-JTL
Timing analysis
& filter
Analysis
results
In-place model
transformation
DesginlevelImplementationlevel
EAST-ADL
design model
EAST-ADL
design MM
C2
u-Rubus
model
RubusMM
C2
C2
u-Rubus model annotated with
analysis results
u-JTL
Rubus
model
Rubus
models
u-RubusMM
C2

12. PROPOSED ENHANCEMENTS
12
C2
Timing
analysis &
filter
u-Rubus
model
RubusMM
C2
u-JTL
Rubus
model
Rubus
models
u-RubusMM
Concretising mechanism. If the leveraged
model-driven techniques could run on
the u-Rubus model, there would be no
need for the individual Rubus models.
Thus, the benefits include fewer activities
and the use of a single modelling artefact.

13. PROPOSED ENHANCEMENTS
13
Analysis
results
Timing analysis &
filter
Analysis
results
RubusMM
C2
Rubus
modelRubus
models
Improved performances of the analysis
mechanism. If timing analysis could run on
the u-Rubus model, the performance would
improve in terms of execution time and
resource usage.
The challenge here is to factorize the timing
analysis with respect to the uncertainty
points.
Alternatively, the analysis engine could
provide estimations for interactively deciding
upon the uncertainty points based on values
of the estimations.

14. PROPOSED ENHANCEMENTS
14
Analysis
results
Timing analysis &
filter
Analysis
results
Exploration chain. Other extra functional
properties (EFPs), e.g., energy
consumption, can affect design decisions.
They can be employed for solving some of
the uncertainty points in the u –Rubus
model resulting from the timing analysis.
The engineer could generate
a new specific u -Rubus model for another
EFP to switch the
focus of the exploration.

15. 15
Model-driven Engineering : Theory and Practice
Track on “Model-driven Engineering: Theory and Practice”at 14th
International Conference on Information Technology : New Generation
WHEN
10th to 12th April 2017
WHERE
Las Vegas, Nevada, USA
Tuscany suites
DEADLINE
14th October 2016
As software systems grow in complexity, traditional code-centric development methods become less efficient.
Modern software systems, in fact, require powerful and flexible development methods able to abstract
problems and solutions as well as automate development phases.
In the midst of the many methodologies advocating abstraction and automation, Model-driven Engineering
CALL FORPAPERS
http://www.mrtc.mdh.se/mdetp
(MDE) has gained major recognition from both researchers and practitioners. MDE focuses on the use of
models for expressing domain specific concepts in a way that is at the same time understandable, precise and
machine processable, thus allowing model manipulations through transformations.
This track aims at providing a discussion forum where researchers and practitioners interested in model-
driven engineering can meet, disseminate and exchange ideas, problems, challenges, identify key issues and
explore possible solutions. We encourage submissions of research papers, experience reports and position
papers, from both academia and industry. Suggested areas of interest include, but are not limited to:
.
• Domain-specific modelling languages
• Model-based analysis, testing, validation and
verification
• Model transformation and reengineering
• Analysis and verification of model transformation
• Model (co)evolution and maintenance
• Model versioning
• Collaborative modelling
• Quality of (meta)models and transformations
• Models@runtime
• Tool support for model-driven engineering
• Model-checking
• Model-based testing
• Model-based measurement, prediction and
monitoring
• Formal methods for MDE
• Integrated tool chains and methods for modelling
and deployment
• Reverse engineering, modelling, and
componentization of legacy code
• Case studies and experience reports
Authors should submit a 6-page version of original and unpublished work including 5 keywords in the
Spinger format. All accepted submissions will be published in the conference proceedings by Springer. If
accepted, one of the authors must attend the ITNG conference and present the work in person.
SPECIAL ISSUE AT JOURNAL OF OBJECT TECHNOLOGY (JOT)
The best papers from the track will be invited for being submitted in extended form to a special issue of JOT.
The extended papers will undergo a new peer-reviewing process
Alessio Bucaioni
Industrial PhD student
at Mälardalen University
alessio.bucaioni@mdh.se
Federico Ciccozzi
Senior lecturer at
Mälardalen University
federico.ciccozzi@mdh.se
Saad Mubeen
Senior lecturer at
Mälardalen University
saad.mubeen@mdh.se

16. Thank you for the
attention!
Questions?