Predictions and decision based on complexly interacting variables are omnipresent in energy, finance and logistics. When using mathematical models to describe such systems, often similar structures are employed like graphs, decision trees or differential equations. Solving these models requires optimisation algorithms or computationally expensive simulations.
The Josef Ressel Centre for Applied Scientific Computing in Energy, Finance and Logistics aims at developing numerical and IT tools for treating such problems.
Within the Josef Ressel Centre we work together with the following Industry partners: Hypo Vorarlberg Bank AG, myPEX, Vorarlberger Kraftwerke AG und Vorarlberger Landesversicherung. Former Partners are: infeo und Gebrüder Weiss.
In vehicle routing, discrete problems with a large number of decision variables have to be solved.
Computationally expensive algorithms depending on a large number of scenarios are omnipresent in finance.
Simulation of power grid requires extensive simulations when considering the dynamic loads of the users.
Market liberalization in combination with an increase in decentralized, fluctuating power generation from renewable sources led to a paradigm shift in energy economics. Demand Side Management (DSM) is considered a tool to adapt the power consumption to such a fluctuating generation. The pursued approach is based on a unidirectional communication of an incentive function, which is created from the operator of the DSM program (e.g. energy utility company) and send to autonomous devices, where it is used to drive the minimization of the energy costs. To optimise directly on the device, the physical behaviour of the load must be modelled and the resulting models must be parameterized. We concentrate on devices with large electrochemical and thermal storage capacities and sufficient power, like hot water storage, stationary battery storage and refrigeration systems. To investigate the effects of the DSM on low voltage grids, a load flow simulation is developed. This simulation allows for the integration of autonomous devices and the implementation of various types of incentive functions.
Balance sheets from banks and insurances contain all relevant information regarding the business and are used for determining performance indicators as required in Basel III or Solvency II. Modelling the positions of the balance sheet allows to predict future developments and stress situations. Another emphasis in this area is on the determination of the risk for various financial derivatives and portfolios of such derivatives. In both areas, the models contain many decision variables and solving requires Monte-Carlo simulations and optimisation algorithms.
In logistics, vehicle routing problems such as the famous "traveling salesman problem" and its variations are a notoriously hard class of optimisation problems. Next to the classical variant, variants with additional constraints (vehicle capacity, time windows for delivery/pick-up, fleet of vehicles) will be considered.
The Josef Ressel Centre aims at developing numerical methods for complex problems and to achieve synergies between thematically diverse but strucutrally similar problems. Our main goals are:
To achieve these goals we will develop a distributed execution framework (DEF) which contains the algorithms for optimisation and simulation as library functions. Within the time period of the project, our aim is to provide the DEF to our partners. The algorithms for simulation and optimisation will be developed in the workpackages related to the application fields.
Josef Ressel Centres are 5-year research projects which are jointly funded by the Federal Ministry Republic of Austria Digital and Economic Affairs (BMWD) and industrial partners. The Christian Doppler Forschungsgesellschaft has the coordination for this Public Private Partnership.