High-Performance Data Analytics Framework for Power Markets Simulation
DXT Commodities is an international energy trading house based in Lugano. Recently, DXT has heavily invested in the development of innovative trading software, including a fully integrated in-house platform covering the whole trading process. On top of this, a lot of effort was spent on implementing forecasting models aimed at supporting trading decisions such as algorithmic techniques by using innovative AI and machine-learning. However, real-time responses of the pricing models, considering realistic engineering constraints encountered in power grid operations, are essential in modern trading environments and are not currently available in DXT trading software. The goal of this Innosuisse project between the Institute of Computing and DXT is to develop computational tools for simulating power market prices that meets the performance requirements of today trading environments. State-of-the-art modeling paradigms of the power grid components will be adapted, taking into account real-life aspects of the problem, the technological constraints of power generators, weather patterns over large geographical areas and environmental constraints. The performance aimed at by the analytics tools lies in the combination of research advancements in terms of numerical methods, mathematical software tools, and in the efficient exploitation of modern computational infrastructures. The USI research group will contribute to the development of novel computational solutions, leveraging its extensive experience and relevant research performed in the domain of efficient solution strategies for large-scale optimization problems in the power grid domain, including problems spanning long time horizons and incorporating renewable energy sources, stochastic problems accounting for their intermittent nature, and considering numerous contingency events in the power grid operations.