Overview
A hybrid quantum-classical tensor network algorithm has been developed to efficiently treat non-linearities within the domain of quantum computational fluid dynamics (QCFD). This approach aims to circumvent challenges posed by implementing non-linear terms on inherently linear quantum hardware, which typically necessitate resource-intensive workarounds limiting scalability. The algorithm combines variational time-stepping with quantum tensor programming to compile operators and time-dependent fields into quantum circuits. Within a probabilistic framework, it replaces prior state-based non-linear implementations with tensor-based block encodings.
Research Context
Non-linear terms constitute a fundamental challenge for quantum computational fluid dynamics. Their implementation on quantum hardware typically requires methods that are resource-intensive, which subsequently limits the scalability of QCFD simulations, particularly for large-scale applications. The linearity of quantum hardware creates a bottleneck when attempting to model the non-linear dynamics inherent in fluid systems.
Approach
The presented algorithm is a hybrid quantum-classical tensor network approach. It integrates two primary components: variational time-stepping and quantum tensor programming. This combination is utilized for the compilation of operators and time-dependent fields into quantum circuits. A key aspect of the methodology involves replacing prior state-based non-linear implementations with tensor-based block encodings. This replacement operates within a probabilistic framework and is designed to stabilize success probabilities, which would otherwise decay exponentially with increasing system size. The algorithm was benchmarked using turbulent flow fields.
Findings
- The hybrid algorithm maintains high success probabilities across increasing Reynolds numbers and grid resolutions when benchmarked on turbulent flow fields.
- It exhibits moderate measurement overhead across increasing Reynolds numbers and grid resolutions during benchmarking.
- The tensor-based block encodings implemented within a probabilistic framework stabilize success probabilities that would otherwise decay exponentially with system size.
- Compared to fully classical tensor network solvers, the hybrid approach yields substantial reductions in memory footprint.
- The hybrid approach also yields substantial reductions in computational cost when compared to fully classical tensor network solvers.
Why This Matters
The development of this hybrid algorithm establishes a scalable pathway toward achieving practical quantum advantage in scale-resolving computational fluid dynamics (CFD) simulations. By efficiently addressing the non-linearity bottleneck, it facilitates more effective utilization of quantum hardware for complex fluid dynamics problems that are currently resource-intensive or intractable with existing quantum approaches.