by Amara Graps
Some of the most visible work by NATO these months are preparations by Eastern Flank countries to build multilayered defenses for what dozens of seasoned military analysts claim will be inevitable attacks on our territories. Then too, there is the less visible work by NATO. One example: enhancing military detection and tracking systems.
There’s a lot of work being done right now on the wakes and the plumes left by hypersonic weapons […]
Computational Fluid Dynamics tools have an important place inside of defense preparations to enhance the performance, survivability, and operational efficiency of a wide range of military systems, from air and sea vehicles to missile defense and battlefield safety.
NATO interest in Quantum CFD
Last July 8-12, 2024, the von Karman Institute for Fluid Dynamics in Belgium with NATO’s support, held a Lecture Series titled: Introduction to Quantum Computing in Fluid Dynamics. It covered topics like quantum algorithms for solving differential equations, hybrid computing for the NISQ era, and quantum machine learning and partial differential equations. It was well received (LinkedIn photos).
From the Workshop’s individual lecture titles, we see again the two quantum algorithm primitives that we discussed in our earlier QCR article about Differential Equations:
- Quantum Fourier Transform (QFT)
- The Harrow–Hassidim–Lloyd (HHL) algorithm
Lattice-Boltzmann and Lattice Gas Automata Methods – European Investment Council
Plus, multiple talks about Lattice-Boltzmann Methods (here is a good classical introduction -pdf).
Those lectures at the Quantum Computing in Fluid Dynamics workshop on Lattice-Boltzmann Methods were why the European Investment Council (EIC) adopted CFD- Quantum algorithms for lattice-based computational fluid dynamics models as one of the most “interesting for further exploration”, in their September 5, 2024 Foresight Report: (Dis)Entangling the Future, which we discussed in the context of AI here. The EIC Foresight exercise chose quantum algorithms for lattice based CFD as a “Novelty” and “promising and interesting” with the following assessment by their external experts:
Computational fluid dynamics (CFD) is a highly computation-intensive field of technology and extremely important for such industries as aerospace, automotive, biomedical, and climate tech, to name a few. Lattice-based CFD algorithms, such as the Lattice Boltzmann Method and Lattice Gas Automata, show high potential for adaptation on quantum computing devices and a clear possibility for quantum advantage. This is an emerging field of quantum computing, as shown by the increasing body of scientific literature around this topic.
Lattice-Boltzmann and Lattice Gas Automata Methods- Why? Quanscient’s Riippi explains
The ‘Why?’ answer for the Lattice Boltzmann Method and Lattice Gas Automata methods is perfectly described in Jousef Murad’s August 20, 2023 Deep Dive #105 podcast, by Juha Riippi, CEO & Cofounder of Quanscient.
Riippi explained, that with CFD, one begins with a multiphysics framework. After a typical equation discretization, you arrive at linear algorithms, such as finite element methods, to solve the partial differential equations. To implement quantum computing at this stage, requires a quantum linear algebra algorithm like HHL. However, for a full implementation, you need fully error-corrected, fault-tolerant quantum devices. Riippi says that the time frame for that is ‘Best estimate: 10 years away”. He said that there is then the Variational approach, which can already be used in the current NISQ era (See QCR’s Part 1, Part 2, Part 3 about hybrid quantum-classical computing).
However, as he explains, the main problem is data input and output. There is no quantum memory, i.e.: quantum RAM. You cannot input an arbitrary matrix into a quantum computer, and you cannot read out the full solution. You can only extract some global properties, such as the integral of the solution.
What Quanscient does for CFD, is put the computing effort into quantum native methods with a focus on algorithms: the Quantum Lattice-Boltzmann (LB) Method and the Quantum Lattice Gas Automata. With these algorithms, one can encode the physics directly on the qubits, [e.g., to map the Hamiltonian of the problem onto the Hamiltonian of the hardware]. With this quantum native approach, there are two advantages:
1) The data input problem is removed because the probabilities of the LB method can be encoded directly onto the qubits of the quantum circuit. The size of the Lattice, or of the problem, exponentially increases, as the number of qubits increases.
2) The data output problem is removed because the qubit measurements correspond directly to the probabilities of the densities in the Lattice. So, the data output from the qubit measurements can be generated to the full solution with a simple summation.
Juha Riippi talked more about the advantages of LB methods at Q2B, 2023, Paris in May 2023. He showed Quanscient’s CFD simulation success from Summer 2022, with a 1D diffusion-advection simulation implemented with Quantinuum and IBM quantum devices. The simulation utilized 7 qubits, yet, with very accurate results. The LB algorithm scales exponentially, and after optimization, scales logarithmically for the number of required gates. With 60-80 effective qubits, he said, they can already solve problems that are pushing the limits of what can be done on classical hardware. He added, that for reference, that would be 580-750 CNOT gates.
Riippi began his Q2B 2023 presentation stating that his company mandate is to show quantum advantage within 2 years. After you look at his results, such a claim looks within reach.
In GQI’s Quantum Computing Focus Report (*):
The race is on to achieve quantum advantage and build practical, commercially viable quantum computers. These investments serve as rocket fuel, propelling hardware players forward as they tackle the immense scientific and engineering challenges involved. Continued support from governments, tech giants, and the broader investor community will be instrumental in maintaining momentum and translating theoretical concepts into real-world breakthroughs.
And further on in the same Report, Quanscient’s mandate would appear to follow the “Broad NISQ” theme (*)
In the coming years (2025-2030), we anticipate three potential scenarios in the Quantum Computing realm. The first and most optimistic scenario involves the realization of Broad NISQ Quantum Advantage, primarily due to advancements in error handling efforts. The second scenario entails the continuation of the current Quantum vs. Classical competition, characterized by claims of supremacy being issued and potentially subsequently debunked by other players in the field or enthusiasts. The third and more pessimistic scenario is the possibility of a Hard Quantum Winter.
(*) If you are interested to learn more of GQI’s highly tuned Outlook Focus Reports, Playbooks, Business trackers, Quantum Technology Frameworks, please don’t hesitate to contact [email protected].
October 11, 2024
