Worldwide, quantum computing in the automotive industry started with quantum annealing and has recently developed in a way that follows the most advanced trends. In this section, we will review the trends in how quantum computers are being used in the automotive industry and introduce them from a technical point of view, based on the examples and technologies we have used in practice so far.
Trend: Consortiums
There is an increasing number of consortia of automobile-related companies in Japan and overseas. In Japan, the Toyota Group and other companies are seeking to utilize this technology, and overseas, BMW and VW are actively participating in consortiums.
BMW and VW to use quantum computers in industry...consortium established
https://response.jp/article/2021/06/14/346688.html
Number of use cases on trapped ion is increasing
BMW is using Honeywell, VW and Aioi Nissay US are using IonQ, and at the cutting edge, the trapped ion with a large number of qubit connections and low errors is gaining popularity. In Japan, quantum annealing and superconducting quantum gates seem to be gaining popularity a little later. The Germans are more sensitive to trends than the Japanese, and they are catching up with the latest technologies, including those of Daimler, so I hope that the Japanese will also do their best.
Quantum chemistry and materials informatics
In the area of materials, battery development is attracting attention, with VW and Daimler taking the lead, and in Japan, TOYOTA is a technology that is attracting attention. The main focus is on quantum chemical calculations using VQE and QPE, and materials informatics using Quantum Machine Learning and other methods.
Supply Chain, Logistics, and Transportation
With the advent of the trapped ion, optimization calculations in fields that were previously performed using quantum annealing are now being performed using QAOA, including supply chain optimization using BMW's Honeywell machine and combinatorial optimization using VW's QAOA optimization by ion trap using QAOA. In Japan, quantum annealing is still being used.
Factory optimization, design optimization, and production optimization
Similarly, various business efficiency improvements using optimization in the manufacturing industry, such as production optimization continue to appear in the news frequently.
Classification problems in quantum machine learning
Aioi Nissay America uses IonQ to perform quantum machine learning using a quantum neural network based on the trapped ion, which is another cutting-edge technology in the field of AI. Although it is not so popular in Japan, we have also presented a similar IonQ circuit for the problem of predicting physical properties at Asahi Kasei Corporation.
Will quantum fluid calculation algorithms develop in the future?
Quantum fluid calculations are beginning to be applied to quantum fluid calculations using quantum amplitude estimation, a technology that is currently being developed mainly in the field of quantum finance. Although the speedup of differential equation-related calculations using quantum amplitude estimation is smaller than that of quantum chemical calculations, which are known as squared acceleration, the field of numerical calculations using quantum amplitude estimation is gradually gaining momentum as the practical application of Monte Carlo calculations in the field of finance is expected to move forward. Fluid calculations such as Navier-Stokes using quantum amplitude estimation are also expected to be applied to car body design in the future.
Four types of software applications for the automotive field are envisioned
First, quantum chemical calculations. Although the speed improvement is unknown, VQE, which uses optimization to solve the Hamiltonian, and quantum chemical calculations using quantum phase estimation, which may be feasible once quantum computers with error correction reach a certain size, are the high road.
Next, combinatorial optimization problems. Currently, we are working on a quantum annealing algorithm called QAOA for NISQ, which can be run on a quantum gate machine. The speedup is severe, but we need to learn various techniques in complex optimization such as constraint equations. It will be interesting to see how QAOA will develop toward FTQC in the future. It is also significant that the structure of all the couplings in the trapped ion makes it possible to perform optimization calculations, which were difficult in superconductivity due to the lack of connections.
Then there is quantum machine learning. It would be good to learn some other techniques for the future.
Finally, quantum amplitude estimation. Using a generalized version of Grover's algorithm, we will learn numerical integration and its application to partial differential equation solving, and explore its future application to quantum fluid dynamics calculations.
Trapped Ion are all the rage in hardware; IBM is also strong.
BMW uses Honeywell, VW uses IonQ. Aioi Nissay US also uses IonQ. In addition to Google and IBM, Amazon has recently been attracting talent at a rapid pace, and Microsoft is also strong. IBM seems to be in the lead when it comes to publishing papers. In terms of technology, Google and others should not be overlooked.
What should we do in the end?
Learn VQE/QPE, learn QAOA, learn QAE, and learn QCBM efficiently. Once you have learned the theory, divide into teams and execute on the actual machine, but be aware that the actual machine is very expensive. That's all.