Hidalgo brings together the HPC, HPDA, and Global Challenges communities in order to bridge the gap between traditional HPC and Data-centric Computation.
HiDALGO focuses on excellent simulations of Global Challenges and in particular, the integration of different problem statements in order to establish a sustainable Centre of Excellence, especially for future problem statements. It advances the proposed technology of the CoeGSS project in any sense, builds on innovative agent-based simulation packages, such as AMOS or RepastHPC and integrates them with the state-of-the-art data analytics tools (e.g. Apache Spark, Apache Flink or Apache Hadoop) and the mighty Vistle many-core visualization platform.
Essentially, HiDALGO relies on coupled simulations, which are common in the engineering sector. By applying a series of dependent simulations at once, distinct problem statements can be analyzed, which represents the only way to address the dependencies of the input parameters. Furthermore, real-world sensor data from the Internet of Things world, such as mobile phones, weather sensors, or even camera data will be integrated in the simulations to improve the statistics-based input data by live information. Taking this approach, statistical errors can be minimized and the accuracy of simulations, especially for rapidly changing influencing factors, significantly increased.
The project is implemented by a consortium of Atos (Spain) – project leader, USTUTT, PSNC, ICCS, BUL, KNOW, SZE, PLUS, ECMWF, MOON, DIA, MK, ARH.
In pilot studies, HiDALGO is using the computing power of HPC systems to solve problems in three selected areas:
– Migration – to predict the destinations of refugees escaping armed conflicts,
– Urban air pollution – a real-time, highly accurate application for urban air pollution simulation and evaluation according to hotspot analysis at street level
– Message spreading in social networks – understanding the structure and dynamics that govern social networks and the processes therein, simulating the behavior of these processes on a large scale and identification of malicious and/or unusual types of postings/messages in these networks.
The main addressees of the project are decision-makers (governments, European Commission), local governments, scientists, and ordinary citizens.