Research projects

Duration: 1.9.2024 - 31.8.2027
Head: Vera Hofer
Funding: EU Commission

Partners: more than 40 partners from industry and research (among them: AVL List GmbH, Mercedes-Benz AG, Infineon Technologies Austria AG, Silicon Austria Labs GmbH, Virtual Vehicle Research GmbH, Graz University of Technology, Technical University of Munich).

One of the project objectives is to investigate the use of cooperative multi-agent systems in rescue operations. Such agents are equipped with simple and low-cost sensors for the detection of heat signals, sound or other signals relevant to rescue operations. During rescue operations such agents may themselves be exposed to risk such that it is unclear whether the agent is able to fulfil its tasks. For example, a drone may get too close to a fire and lose some or all its functions. The University of Graz is researching the problems arising from such situations of risk. Particularly, we are investigating which data and calculation methods are required to determine the remaining probability of success that an agent will fulfil its tasks. This information will contribute to a suitable decision-making process and is required to optimise the execution of critical missions.
 

Duration: 01.10.2020 - 31.12.2024
Head: Ulrich Pferschy
Funding: University of Graz
Project partner: University of Udine

Content:

When studying to become a teacher at the University of Graz, students can choose any two subjects from 28 possible degree programmes. This wide range of choices inevitably leads to overlapping courses and thus to obstacles in the course of study. Using an optimisation model (ILP), the timetables previously planned locally by individual institutes are now to be planned centrally. The number of parallel groups on offer (at least for the subjects that are in greater demand) should enable a largely overlap-free course programme to be calculated that is tailored to the individual student. The usual deviations from the standard progression of studies (deferral of individual courses) are also taken into account.

Duration: 01.01.2018 - 31.12.2020

Head: Vera Hofer

Funding: Austrian Federal Ministry of Science, Research and Economy

Partner: Geological Survey of Austria

With its research of regenerative deposits, the project is dedicated to the social challenge of resource scarcity. It is therefore directly related to the goal of sustainable development as regards the protection and sustainable use of terrestrial resources. As part of the project, techniques for the automation of petrographic scree analysis were further developed using statistical methods based on reflectance spectra of Austrian scree samples. The analysis of such data poses a challenge to the statistical methods to be used due to the volume and structure of the data.

Duration: 01.08.2016 - 30.10.2019

Management: Vera Hofer

Funding: Anniversary Fund of the Oesterreichische Nationalbank (OeNB)

In a dynamic environment the distribution of explanatory variables and of the target variable change over the course of time. Such changes result in a deterioration of the classification rule since the data which a statistical classification model is based on is no longer representative after such changes. 
In many applications it is impossible to obtain a recent and complete dataset including explanatory variables and the target variable. A typical example is the lending business. The explanatory features such as income, occupation, family situation, assets, etc. can be observed easily. However, the customers’ willingness to repay a loan cannot be observed. Such information is available much later. The project explores how a statistical classification model can be updated by means of incomplete data comprising only of the recent explanatory features which provide information on a changing environment, while no information is given about the target variable and its change over the course of time.

Duration: 01.01.2022 - 31.12.2023
Head: Ulrich Pferschy
Funding: Zukunftsfonds Land Steiermark
Project partner: FH Joanneum

Content:

Since 2021, the Renewable Energy Expansion Act has enabled the establishment of renewable energy communities (EEG). This means that locally generated energy, for example from photovoltaic systems, can be produced, stored, consumed and sold to each other within the energy community. The specific rules of the game and pricing within the community can be freely organised. For the individual participants, there are different strategies for dealing with the self-generated electricity. The project will create an optimisation model that can be used to optimise the behaviour of individual prosumers and the community as a whole. Storage elements, time-shiftable consumers and electric cars are particularly important here. Based on the optimal decisions, adequate price models will be developed to guide even selfishly acting members towards behaviour that makes sense for the community as a whole.

Duration: 01.10.2019 - 31.12.2023
Head: Ulrich Pferschy
Funding: FFG / klima+energie fonds
Project partners: Holding Graz AG and others

Content:

Holding Graz plans to operate the city's bus fleet with emission-free drive technologies by 2030 in order to reduce greenhouse gas emissions, air pollution and noise pollution in the urban environment. The move2zero cooperation project is developing a holistic concept for the complete decarbonisation of an urban bus transport system. The drive concepts being considered are electric drives powered by short or long-term batteries, fuel cell drives and hydrogen engines. As a project partner, the institute is developing optimisation models that select the optimal drive concepts for the individual bus routes for given timetable and capacity specifications in order to optimise the overall costs of the transport system. The required infrastructure (e.g. charging stations) must also be planned and optimally localised for this purpose.

Duration: 01.09.2020 - 31.12.2022
Management: Ulrich Pferschy
Project partner: Red Cross Styria

Content:

The number of Red Cross vehicles deployed daily and the personnel required for this in each district centre is a key factor in the satisfactory provision of services and the costs incurred. A forecasting tool is therefore to be developed which estimates the number of deployments for the following day on the basis of the available registrations and other influencing factors such as weekdays, public holidays, seasonal aspects, etc. The optimisation system developed should then suggest the appropriate number and type of vehicles using the deployment data from the past.

Duration: 1 May 2015 - 30 April 2018

Head: Vera Hofer

Funding: EU Commission/FFG

Partners: 39 partners from industry and research (among them: Infineon Technologies Austria AG, Plansee SE, Fronius International GmbH, ams AG, Max Planck Institut für Eisenforschung GmbH, Technische Universität Dresden, greenpower technologies, University of Bristol).

Technical properties of electronic components in vehicles are subject to certain changes over their lifetime. In order to obtain data on the drift behaviour of these parameters, appropriate stress tests are carried out on a random sample of such components. In the project PowerBase the University of Graz focused on the development of statistical and probabilistic models to determine test limits for technical parameters of electronic components based on these data. These test limits are intended to ensure that the components meet the specifications required by the user over the components’ lifetime.