Quality Software Engineering (QSE) Research
Project title: Complex Systems Design &
Engineering (CSDE) – Simulation Production automation
Project period: 1.2.2008 – 31.1.2010
Abstract: Information and communication technology
(ICT) has been successfully used to control and manage complex systems in many
domains. However, decision makers, who
deal with complex systems problems and are not ICT experts, are often not aware
of the range of contributions of informatics research they could use to address
their complex systems challenges effectively and efficiently. The Complex
Systems Design & Engineering Lab (CSDE-Lab) is named after the TUW
Informatics faculty research focus „Complex Systems“. The objectives are to
engineer and manage complex systems in concrete mission critical applications
in safety-critical domains, like mission coordination for emergencies, Air
Traffic Management, or Production Automation.
Characteristics of complex systems are to be difficult to model, predict, and
manage due to their variety, dynamics and emergent system properties.
Additionally, the integration of heterogeneous systems is essential in order to
achieve common goals more effectively, efficiently, and/or more robust to
failures. Thus, the CSDE-Lab develops, evaluates, and improves processes, methods,
and tools for systems development, integration, and verification.
Particular research topics in the CSDE-Lab are: a) the usage of Ontologies as
continuous software engineering model for the specification of technology
capabilities and capacities, support of iterative system reconfiguration and
optimization; b) optimization of coordination patterns with multi-criteria
objectives using space-based computing for efficient production scheduling and
backup solutions with minimal disruption; and c) formal specifications and
strong QA for the verification of mission critical elements, like test
automations to lower the effort for re-testing software systems, the
measurement of system performance, or the checking of assertions at run time to
inform the responsible roles.
People involved: Stefan Biffl, Thomas Moser,
Richard Mordinyi, Eva Kühn, Alexander Schatten
title: Test-Driven Automation
in the logi.cals system environment automation
TDA – logi.DIAG
type: FFG national funding
partners: logi.cals, Messfeld
period: 01.09.2008 – 31.12.2010
"Test-Driven Automation" is supposed to support automation
technicians with a new integrated development process which should help them to
deal with the increasing complexity of extensive automation projects based on
IEC 61131-3 and IEC 61499. During the whole lifetime of the plant the engineer
should be able to use a framework based on test driven methods (from business
IT state of the arte) which raises the outcomes quality while concurrently
decreasing work amount.
Through adapted quality assurance
methods like unit- and module tests based on a new reference architecture in
conjunction with requirements management, reusability and development of SW
product lines are supported. The development process is stabilized for the
implementation of later requirement changes too. With the help of this
architecture a sum able connect of automation and test functions with Diagnose
and Condition Monitoring (CM) components is also possible. The plant monitoring
is considered to be a largely observing test during the running plant. For test
evaluation (diagnose and CM-tasks) appropriate data analysis methods are
developed which are used by automation and service technician as well as for
processing on particular platforms. Test-Driven Automation should fulfill the
number of errors found during the engineering phase and so reduction of found
errors in the integration phase by 30%
of the start-up time by at least 10%
of the diagnose and CM creation time by at least 35%
of errors caused by the chance of requirements by 30%
results: Expected project outcomes are SW-prototypes as well as domain specific
implementations in the area of the roll mill technology.
involved: Stefan Biffl, Dietmar Winkler, Alexander Schatten, Thomas Östreicher
Project title: System-Wide Information Sharing
Project type: FFG national funding; industry
Project period: 01.07.2006 - 31.12.2008
- Frequentis, Austro Control
- Research group "space-based computing" (head: eva Kühn, TU Wien)
Contact: Stefan Biffl.
The research project System-wide Information
Sharing (SWIS) is targeted as an information sharing network within the Air
Traffic Management (ATM) domain, with very demanding safety and security
requirements as well as the need for high availability. Today companies and
organizations operate in a highly complex environment requiring well-defined
but flexible means for communication and cooperation that can be easily adapted
to potentially frequently changing business processes.
Traditionally most organizations have developed
IT infrastructures consisting of numerous stand-alone applications, which are
connected via point-to-point links and thus lack of flexibility. Over the last
years several approaches have been taken to solve this architecture problem,
like Enterprise Application Integration as a concept and the Service Oriented
Architecture. However, these approaches provide only mechanisms for a flexible
interconnection of various business applications in one domain. In the ATM case,
many actors are involved, e.g. airports, airlines, military users, General
Aviation, Air Traffic Service Providers, Air Traffic Flow Management instances,
resulting in more or less de-coupled actions and decisions. However, due to the
expected growth of air traffic in the next decades all ATM actors will be
forced to co-operative handling of virtually shared information during the
entire life cycle of a flight.
To support high-level cooperation, a low-level
mechanism for information sharing needs to be established and the corresponding
operational procedures and practices agreed and installed at all actors’ premises.
In the ATM environment, the degree of heterogeneity of existing legacy systems,
solutions, actors, their practices, and preferences may well preclude any
“end-to-end” interoperable solution. Therefore, it is essential to keep
low-level information sharing mechanisms strictly de-coupled from high-level
applications that rely upon these mechanisms.
The demand for an improved solution in the ATM
domain generates a need for the development of a “System-wide Information
Sharing” (SWIS) based upon adequate and sound concepts. Major contributions are
a) the description of the source and target systems in terms of data
requirements and functional ranges as well as constraints of the infrastructure;
b) algorithms deducing “intelligent plumbing” from semantic descriptions and
defining an automated way to acquire the canonical data exchange model as well
as the resulting data flows and interface function calls; c) the description of
the solution architecture with all functional components; and d) the development
of a prototype integration for example applications utilizing the results of
the research project. SWIS enables sharing of information in a highly
distributed environment, taking demanding requirements regarding performance,
scalability, maintainability, safety and security into account.
Amin, BIFFL Stefan,
MOR Marcus, MORDINYI
Richard, MOSER Thomas
Project title: Ambulance Routing – Fleet
management for Emergency Service Vehicles
Project type: Austrian Science Fund (FWF) national funding for translational research; academic partner University of Vienna
Project period: 01.01.2006 - 31.12.2008
- University of Vienna, Department "Production and Logistics" (Prof. Hartl)
Contact: Stefan Biffl.
emergency service providers, especially ambulance departments and companies who
provide non-public maintenance services, face the problem to provide different
types of services with one fleet of vehicles:
Emergency coverage for a certain region to provide immediate emergency service;
Efficient regular service: scheduled pick-up and delivery of patients,
predetermined service tasks, periodic pick-ups, etc.
is also the current situation for the largest Austrian regional emergency service
providers (e.g., the Austrian Red Cross), where the same fleet is used to
provide both emergency and regular transport services. Dynamic emergency
aspects thus directly influence the schedule for the regular service. When an
emergency occurs and an ambulance is required, the vehicle with the shortest distance
to the emergency is assigned to serve the emergency patient. Therefore, it
often happens that an ambulance vehicle that has been scheduled for a transport
order of a patient, but has not yet started, serves the emergency request.
Thus, another vehicle has to be reassigned to the regular patient and the
overall regular service schedule has to be re-optimized. Ambulances that carry
out emergency transports become available at the hospital after the emergency
service and can then be used to carry out regular transportation orders. Again,
the schedule for regular services has to be re-optimized.
transportation services are offered for handicapped persons or patients with
minor diseases, who could not use taxi services. Thinking of optimization for
ambulance scheduling, we have to consider at least two perspectives. From the
perspective of a transportation provider, the objective is to minimize cost of
operations. On the other hand, to maximize quality of service is the objective
from a patient’s point of view. Although both objectives comprise a multitude
of factors, a simplified model of reality is subject to our investigations.
Basically, we use the length of a tour in terms of driving time to model costs
and waiting time of patients to model transportation quality. A tour or route
is defined as the overall movements of a vehicle over a day of operation.
project dealt with different perspectives on the ambulance scheduling, e.g.
consideration of expected transportation requests in vehicle scheduling
determined from experiences in the past or waiting strategies for maximizing
coverage to reduce response times for emergency services. In this contribution
we will concentrate on regular transportation services, where a minor part of
transportation requests arises dynamically and most of the requests are known
beforehand. Emergency requests disturb regular operations, but may be modeled
as dynamic requests with high priority.
the constraints in classical Dial-a-Ride Problems we have the following
features of our problem:
Different (hard as well as soft) time constraint types; e.g., dialysis patients
must arrive exactly on time because the dialysis machine is reserved.
Heterogeneous fleet - We deal with a heterogeneous fleet concerning capacities
of the vehicles and concerning the different available equipment on the
Dynamic aspects - Some orders are known in advance. Additionally, the vehicle
availability changes dynamically. The reason is the disappearance and
reappearance of vehicles (emergency requests are serviced with the same fleet).
Stochastic aspects - Expected return transport orders (relatively long planning
horizon), expected availability of vehicles, expected transport orders
(relatively short planning horizon, in some cases emergencies occur and most
probably additional vehicles are required.)
major goal of the Ambulance Routing project is to demonstrate potential
advantages of optimization algorithms in a decision support pilot system for
ambulance scheduling. Related work in the field of decision support systems has
demonstrated the practical use of optimization for real-world vehicle routing
problems. Providing dynamic routing services requires a certain information
system infrastructure that integrates Positioning Systems, Wireless
Communication, and Geographic Information Systems to process necessary inputs
for optimization and decision support, which in turn can provide business
services useful for dispatchers.
contributions of this project are twofold. On the one hand, we describe two
promising and efficient solution procedures suitable for a bi-objective version
of the dynamic “dial-a-ride problem” (DARP) and evaluate the efficiency of
these procedures with real-world problem instances.
the other hand, we review restrictions and constraints for the development of a
decision support extension for ambulance scheduling and suggest a system
architecture for integration of our solution procedures into an existing
control center system to provide interfaces for future business services.
Project results: http://csde.ifs.tuwien.ac.at/csde/csde/ambulanceRoute
People involved: BIFFL Stefan,
Czech IT Project Process Improvement for Dependable-Systems Research based on the Novel V-Model XT Approach
- Czech Technical University, Department of Cybernetics (Prof. Marik)
- Sponsored by EU Marie Curie Transfer of Knowledge Program
Contact: Stefan Biffl.
The Gerstner Laboratory (GL) at the Czech Technical University in
Prague successfully conducts research in the areas of artificial
intelligence, with the focus on knowledge-based and information
systems, knowledge discovery in databases, multi-agent systems, and
software diagnostics. Till now GL has gathered experience in the
development of many IT systems in the above mentioned areas. Due to an
extreme increase in the size and complexity of software development in
research projects the ad-hoc software process has already reached edge
of quality and economics.
At this stage it becomes
necessary for GL to build up competence in the field of "software
process improvement for dependable systems" in order to enhance the
research conducted in areas of dependable systems applications. This
new competency will enable GL researchers to overcome the current
challenges in the IT projects management (high risk of rework, risk of
feasibility, quality and usability of end product and services) and to
support GL researchers with the advanced methodologies addressing the
quality assurance, quality management and standardization in software
The specific knowledge transfer objectives are formulated as follows:
- To acquire a knowledge of formal
methodology of software development process, software engineering,
software process improvement,
- To learn the concepts of V-Model XT
framework and to tailor its process models to the needs of GL research
projects and application domains in Czech practice (aimed at dependable
- To integrate methods from requirements management and knowledge engineering to refine and improve software process models,
- To enable researches to conduct
empirical studies in the area of software engineering and quality
management to improve processes and products.
The partner organization
providing training for the outgoing researchers is TUW, The Institute
of Software Technology and Interactive Systems (IFS). IFS has a
considerable experience in Quality Software Engineering (QSE) Research,
teaching and training in the area of software engineering, project
management, quality management, and risk management. Additionally IFS
will provide the training in the V-Model XT framework for the GL
Requirements Management in Global Software Development - Using Traces to enable Requirements Awareness and Information Drill-Down
Contact: Matthias Heindl.
The main goal of software and systems engineering projects is to
fulfill customer requirements, e.g., to deliver software according to
the stakeholder value propositions.
The increasing complexity of
functionalities, the multiplicity of stakeholders, shorter
time-to-market, outsourcing, and highly distributed project teams of
today's projects makes the following requirements management tasks
difficult for project managers, requirements engineers, and other
- Explain the range of stakeholders' value
propositions for the desired software (or system) to all project
participants and obtain a common understanding of requirements; This is
a challenge especially in global software development (GSD) projects
because of the limited possibility for clarification with the
- Understand a given task in a GSD project
setting: This is a problem of information provision: e.g., the
developers who get a new task from their boss usually do not get enough
context information. This situation results in lots of telephone calls
and email threads involving multiple people to clarify tasks.
My research deals with
requirements management methods, concepts, and tools. The focus is on
requirements tracing and applications of traces. Requirements tracing
is the ability to follow the life of a requirement both in a forward
and backward direction, e.g., by capturing and maintaining
interdependencies between requirements and other artifacts emerging
Generally, traces are a means to
create relationships between pieces of information that belong together
in a certain way. For instance, value propositions of stakeholders can
be traced to user requirements and their rationale, which in turn can
be traced to the components that implement these requirements. Such a
chain of pieces of information make it easier to understand the origin
of tasks and requirements, and thereby support developers in
implementing tasks correctly and completely.
The research issues in this context are:
Research Issue 1: Cost-benefit optimization for creation and maintenance of traces by using tools and value-based approaches;
Usually, tracing is done manually, which is expensive and error-prone.
Combining value-based approaches (capturing only the valuable traces)
with automation (to reduce effort and errors) is a promising attempt
Research Issue 2: Exploring new applications of traces;
As mentioned above, traces are a means to relate pieces of information.
Thus, traces could be used to provide project members of GSD projects
with the context information they need to fulfil a task effectively and
efficiently without the need for long-winded email-threads or telephone
calls. Instead, a trace-based information provision tool (lets say a
"network bench") should allow for each project participant to perform
an "information drill down": Each team member should be provided with
the relevant information he needs for his current task without being
spammed with unnecessary information. An "information drill down"
allows him to retrieve the following context information for his task,
which can be related by traces:
- A task description (To Do list instead of intuitive hacking)
- A request that the task is derived from, e.g., a requirement that has to be implemented
- related requirement information and artefacts
- history of activities that were already performed in context of the new task
- relevant contact persons
- communication (emails) and discussions that already took place in context of the new task (history)
The results of the work will be:
(a) case studies that illustrate the improved cost-benefit of
value-based tracing approaches and (b) a concept and prototype of a
project platform that improves efficiency and effectiveness of the
project members in solving their tasks. The trace-based project
platform provides them with the necessary context information for their
task so that long-winded email threads or telephone calls for
clarification can be avoided.
'Software Process and Product Improvement with Inspection'
Contact: Stefan Biffl.
management and quality management are core competencies for timely and
economical development of high-quality software products. This project
focuses on the following three inspection-related areas: Defect
Detection, Defect Content Estimation, and Reinspection. The project
consists of designing and conducting a series of large-scale controlled
experiments to address the above-mentioned research issues.
Computer-supported Inspection Methods in Software Engineering for small software developers
Contact: Michael Halling, Paul Grünbacher.
Classical (manual) inspections show some inefficiencies which probably
can be overcome by modern information technology, e.g. networked
computers allowing decentralized and asynchronous work. At the same
time there are indications that the "one-or-all" approach to
inspections is not optimal. Diversification of inspections dependent on
the type of project, the domain and the culture promises even better
results. In this project we
- Investigate the current state of computer supported inspections.
- Compare these methods with respect to their underlying model.
- Analyze these methods with respect to
their applicability in various project contexts, especially for small
software developers (SSDs).
- Build an own prototype using an Electronic Meeting Room approach.
- Investigate the effects of these modern
techniques on the synergetic effects of face-to-face inspections
meeting claimed by Fagan.
- Provide a 'road map' advising users of decision criteria for various inspection types.
- Develop a handbook plus CD-ROM containing useful and practical approaches, forms and examples, etc.
- Validate our findings and our deliverables with practitioners from the field.
- Disseminate our findings in conferences and scientific journals.