Back
Qubit-aware quantum algorithm development for industrial use cases
A challenging aspect in many computational engineering applications is that physics-based forward simulations have to be embedded into high-throughput tasks to enable sensitivity analyses, uncertainty management, optimization or parameter estimation tasks. Such secondary analyses can easily become computationally impractical, requiring a large number of simulation runs. Gaussian process emulation, also known as Gaussian process regression is a well-investigated technique to train a fast to be evaluated surrogate model for such tasks. An advantage of the GP based surrogate is given by its practically useful error bounds. However, GP emulators are currently limited to a moderate dimension of the design space, as the training of a high-dimensional Gauss process itself is computationally infeasible. Recently, quantum alternatives have been suggested to tackle this limitation in the future. This project explores the suggested approach further, with a particular focus on potential and relevance for computational engineering applications.
Field of action:
Quantum Computing
- Faculty 4 – Mechanical Engineering
Address:
Building: ESS, Room 210, Eilfschonsteinstraße 18, 52062 Aachen
Contact:
Dr. Sc. habil. Julia Kowalski
kowalski@mbd.rwth-aachen.de
Homepage:
https://www.mbd.rwth-aachen.de/cms/mbd/der-lehrstuhl/team/~qashd/julia-kowalski/
Status:
running