Accelerating Simulations using Hybrid Quantum-Classical Machine Learning

Hamzeh Kraus & Rahul Banerjee

Simulations play a crucial role in the development process of vehicles in the automotive industry. One of the most challenging aspects of these simulations, which can take hours or even days, is reducing the runtimes. Machine learning methods are capable of addressing this challenge. In particular, the deployment of neural networks can reduce runtimes to a few minutes, as they are able to take advantage of the tremendous parallelization capabilities of GPUs. The question is: Is it possible to shorten the runtime even further? Indeed, theoretical investigations of quantum computing have demonstrated that it can accelerate the solution of several problems by orders of magnitude. In this study, we apply Quantum Machine Learning to predict the suspension oscillations of a passenger car experiencing excitation of its spring-damper system on a rough, bumpy road. Our application, which can run on any CPU, GPU, or quantum computing backend, is used to examine both the quantitative and qualitative features of the final product.

To the top of the page