Tutorials
The role of the tutorials is to provide a platform for a more intensive scientific exchange amongst researchers interested in a particular topic and as a meeting point for the community. Tutorials complement the depth-oriented technical sessions by providing participants with broad overviews of emerging fields. A tutorial can be scheduled for 1.5 or 3 hours.
Simulation and Verification of Cyber-physical Systems
Instructor
|
Hermann Kaindl
TU Wien
Austria
|
|
|
|
Brief Bio
Hermann Kaindl joined the Vienna University of Technology in early 2003 as a full professor. In the same year, he was elected as a member of the University Senate. Prior to moving to academia, he was a senior consultant with the division of program and systems engineering at Siemens AG Austria. There he has gained more than 24 years of industrial experience in software development. His current research interests include software and systems engineering focusing on requirements engineering and architecting, and human-computer interaction as it relates to interaction design and automated generation of user interfaces. He has published 5 books and more than 240 refereed papers in journals, books and conference proceedings. He is a Senior Member of the IEEE, a Distinguished Scientist member of the ACM, and he is on the executive board of the Austrian Society for Artificial Intelligence.
|
Abstract
Automotive systems, for instance, are safety-critical cyber-physical systems (CPS). In particular, undesired feature interaction can lead to safety-critical behavior. In order to address this problem, we investigated physical feature interaction in this context using simulation (with more than one physical variable). This allowed us to visualize both the behavior of features in isolation and their interaction. For studying the effects of deviations for uncertain inputs of systems, often multi-run simulation is employed, which is time-consuming. Unfortunately, such simulations also do not directly support the traceability of such effects. A semi-symbolic modeling approach based on Affine Arithmetic Forms (AAF) allows the representation of uncertainty in terms of ranges. Simulations of such models directly include propagation of deviations and their traceability. We studied such a semi-symbolic model of a CPS, including coordination of safety-critical and interacting features. For feature coordination, this model introduces handling discrete uncertainty with two different behavioral modes and their integration. In addition, we investigated coordination of physical feature interactions using model checking. In particular, we created and used a qualitative model for formal verification against a property in time logic. This model is intended to be minimalist, in particular the logical model based on a physical model (including speed and distance). This logical model defines the essence of operations in the dedicated environment. We also investigated formal verification of a specification of a software coordinator using the Fluent Calculus (a derivative of the Situation Calculus). Based on this research, this tutorial also shows that and how formal verification of (safety-critical) CPS can be made in addition to simulations.
Keywords
Affine Arithmetic Forms, model checking, Fluent Calculus
Aims and Learning Objectives
This tutorial has the primary objective to provide the participants with a better understanding of simulation, both traditionally and using AAF, and of the difference to formal verification.
Target Audience
The target audience is practitioners, students and educators.
Prerequisite Knowledge of Audience
Attendees are supposed to have some basic familiarity with simulation. But they do not need to know already about formal verification, which will be introduced in this tutorial intuitively.
Detailed Outline
5min Introduction
20min Background
- Traditional Simulation
- Semi-symbolic Simulation using Affine Arithmetic Forms
- Formal Verification
30min Simulation of Physical Feature Interaction
- Modeling of Cyber-physical Systems (CPS)
- Simulation Approach
- Simulation and Visualization of Results
40min Semi-symbolic Simulation and Analysis
- Modeling using Affine Arithmetic Forms
- Connecting Based on Semantic Task Specification
- Minimum Coordinator Block
- Semi-symbolic Simulation
- Analyses
40min Minimalist Qualitative Models for Model Checking
- Modeling Approach
- Qualitative Model
- Model Checking Results
40min Verification of Feature Coordination Using the Fluent Calculus
- Modeling in the Fluent Calculus
- Verification Based on the Fluent Calculus
- Contrasting With Model Checking
5min Summary and Conclusion
Software Quality Predictive Modeling using Machine Learning: Present and Future
Instructor
|
Ruchika Malhotra
Delhi Technological University
India
|
|
|
|
Brief Bio
Dr. Ruchika Malhotra is Associate Head and Associate Professor in the Discipline of Software Engineering, Department of Computer Science & Engineering, Delhi Technological University (formerly Delhi College of Engineering), Delhi, India. She is Associate Dean in Industrial Research and Development, Delhi Technological University. She was awarded with prestigious Raman Fellowship for pursuing Postdoctoral research in Indiana University Purdue University Indianapolis USA. She received her master's and doctorate degree in software engineering from the University School of Information Technology, Guru Gobind Singh Indraprastha University, Delhi, India. She was an Assistant Professor at the University School of Information Technology, Guru Gobind Singh Indraprastha University, Delhi, India. She has received IBM Faculty Award 2013. She is recipient of Commendable Research Award by Delhi Technological University. Her h-index is 25 as reported by Google Scholar. She is author of book titled “Empirical Research in Software Engineering” published by CRC press and co-author of a book on Object Oriented Software Engineering published by PHI Learning. Her research interests are in software testing, improving software quality, statistical and adaptive prediction models, software metrics and the definition and validation of software metrics. She has published more than 150 research papers in international journals and conferences.
|
Abstract
Software quality is defined as the ability of a software to meet its stipulated requirements and achieve customer satisfaction. Practitioners and researchers all over the world are actively devising methods which ensure a good quality software. One promising approach for doing so is the use of software quality predictive modeling. It involves development of models for estimating software quality attributes such as maintainability, defect-proneness, reliability and others in the early phases of software development life cycle. With the aid of such models, software managers are capable of identifying high risk program modules, which can be rigorously monitored and tested to ensure a successful software product. Such models also assist in efficient use of constraint software resources as these resources may be allocated to modules, which are more prone to defects and changes. Such practices assure a low cost, reliable and maintainable software product, with a highly satisfied customer base. Researchers have demonstrated the effectiveness of machine learning in developing effective software quality prediction models. Machine learning simplifies the learning process as it has the ability to automatically adapt and learn from previous experience and historical data. As machine leaning techniques do not need to be explicitly programmed, they are robust and adapt easily to changing circumstances. We discuss the step by step process of software quality predictive modeling using machine learning techniques so that one may conduct effective empirical studies, with the aim to develop successful predictive models. We present the current status and also throw light on the future avenues in the field in order to guide avid researchers interested in the field.
Keywords
Predictive Modeling, Software Quality, Machine Learning
Aims and Learning Objectives
• To provide details of existing methods and techniques in the area of software quality predictive modeling.
• To provide future guidelines for applying machine learning techniques in software quality predictive modeling.
Target Audience
The tutorial is targeted at academic researchers and software practitioners who plan to investigate and empirically validate the relationship between various software quality attributes and OO metrics. It presents and proposes a definite research procedure to be followed while doing replicated empirical studies with Object Oriented metrics as the independent variable and a particular software quality attribute (fault proneness, maintenance effort or testing effort) as the dependent variable.
Prerequisite Knowledge of Audience
Basic knowledge of software metrics and quality attributes
Detailed Outline
Part I: Basics of Predictive Modeling
What is Predictive Modeling?
Steps in Predictive Modeling
Key Elements in Predictive Modeling
Part II: Research Methodology for Software Quality Predictive Modeling (SQPM)
Empirical Study Process
Study Definition
Experiment Design
Research Conduct and Analysis
Results Interpretation
Reporting
Part III: Research Issues
What kind of repositories are available for extracting software engineering data?
What are the various software engineering repositories and open research data sets for SQPM?
What type of data pre-processing and feature selection techniques should be used before developing predictive models?
Which possible tools are freely available for mining and analysis of data for developing software quality predictive models?
Which techniques are available for developing software quality predictive models?
Which metrics should be used for performance evaluation of SQPM?
How can we effectively use search-based techniques for predictive modeling?
Which statistical tests can be effectively used for hypothesis testing using search-based techniques?
Part IV: Current Trends in SQPM
Part V: Future Directions of SQPM