BMW Selects Honeywell H1 Quantum Computer and Entropica Labs Model for Quantum Demonstration Chain Supply Chain Proof

BMW carries over two million cars every year. Approximately 30,000 parts in each car are supplied by more than 100 global suppliers. It is easy to see why global supply chain management and coordination is a complex and computerized process. BMW, like other manufacturers, is currently limited to managing the supply chain and logic with classic computers and software. While quantum computing is still at the prototype stage, BMW is interested in seeing if quantum computing has the potential to maximize and accelerate supply chain management. In the long run, the hybrid use of quantum computing and artificial intelligence will be a powerful tool for managing consumer demands, political movements, tariff changes and other environmental impacts.

For this project, BMW teamed up with Entropica Labs, a Singapore-based startup computing company, and Honeywell Quantum Solutions. It was thought that developing and running appropriate criteria for near-term quantum computers could provide BMW valuable insights on the value of quantum capability.

Honeywell’s H1 quantum computer uses an advanced lock-ion architecture strong enough to support the next generations of Honeywell quantum processors. H1 computing power comes from ten fully bonded ytterbium ions and features high fidelity activity and low crosstalk. Honeywell currently holds the highest quantum size record of 512 for quantum computing.

Honeywell was the first to develop a mid-cycle measurement and reset and conditional feedback (MCMR) feature.

Entropica Labs

Entropica was selected to create the necessary quantum algorithms and run them on the Honeywell H1yw Model. Entropica was founded in 2018 by Tommaso Demarie and Ewan Munro, an alumni of the Quantum Technologies Center (CQT) in Singapore. In 2020 it raised 1.9 million USD in a seed funding round led by Elev8, a VC company with a focus on high-tech opportunities. In addition to prior experience with Model 0 and Model H1 Honeywell, Entropica has also worked with other quantum cloud providers, including Rigetti Computing, IBM and Microsoft.

Cofounder Ewan Munro is the CTO of Entropica. He holds a Masters in Mathematical Knowledge from the University of Edinburgh and a Ph.D. in Physics from the Center for Quantum Technologies at the National University of Singapore.

Tommaso Demarie, Head of Entropica, studied physics in Italy before moving to Australia to complete his Ph.D. in quantum information theory at Macquarie University in Sydney. He later became a Postdoctoral Research Fellow at the Singapore University of Technology and Design and the CQT.

Munro explained that Entropica had previously conducted research using Honeywell’s first quantum computer, the H0 model, as well as the first release of the Model H1. Entropica’s previous Model H1 research focused on examining mid-cycle measurement and repositioning (MCMR) capabilities. Because MCMR allows the interpolation re-use of squares, Entropica created a machine learning application that cleverly used 6-qubits to run an application of “Bars and Stripes” that required 9-qubits. The three additional qubits were “parasite resources” recovered from the use of MCMR.

Munro said, “That was a very interesting problem, and the results were great.”

In another example of how powerful this feature is, I went to a recent quantum conference where Honeywell demonstrated a similar application to mid-cycle measurement and repositioning using an algorithm called Bernstein-Vazirani.

In plain English, Bernstein-Vazirani can decipher a secret string that is hidden in another action. Assume a 6-bit string is hidden in a file. The goal is to determine the value of the hidden number and find it in the least number of attempts. A classic computer had to make six estimates to determine the secret number. The beauty of Bernstein-Vazirani is that it allows a quantum computer to find the right answer in one attempt.

For the demonstration, researcher Honeywell first ran the Berstein-Vazirani quantum algorithm 6-qubits to find the hidden value. The next show showed that he was running just as well as a 2-qubit a tour that introduced MCMR.

Benchmarks

To validate BMW, Entropica used a mathematical protocol called number division. It is a classic organizational problem that provides a common entry point for many logistics and supply chain problems of business interest.

For the most part, today’s quantum computers cannot do anything that cannot be done on a classic computer. To determine the accuracy of quantum tests, researchers routinely run the same or similar algorithms on classical simulators or classical computers as a benchmark.

Entropica Runs the Quantum Recursive Estimation Optimization Algorithm (R-QAOA) on the Honeywell H1 Model. To compare H1 results, the R-QAOA algorithm was also run on a classical simulator. For a classic computer benchmark, Entropica selected the Karmarkar-Karp algorithm. It is the classic heuristic for number division.

Simply put, the proof of concept for BMW works:

The algorithms divide a set of positive numbers into two groups called divisions, starting with the two largest numbers and ending with the smallest numbers. The “fixed difference” is the total difference between the sums of the two separations. The smaller the set difference, the better the result.

How does this conceptual test apply to logistics and supply chains?

This exercise could be an example of loading material onto two different vehicles. In such a case, the goal is to ensure that each vehicle does not have more weight than the other. It could also represent a loading rate between two servers.

Results

The table shows set differences for each approach. The x-axis variable is the magnitude of the problem, while the y-axis variable measures performance. It is easy to see that R-QAOA running on the Honeywell H1 Model compares favorably with R-QAOA running on the classic bullet-based simulator (denoted by ‘sim’). This indicates that there was little noise from the H1 device itself, compared to the variables due to the finite number of measurements taken.

For the small problem scenarios studied on the algorithmic side, the depth-1 R-QAOA has comparative performance compared to the classic heuristic Karmarkar-Karp (KK).

Commenting on Honeywell’s findings, Demarie said, “At this stage of the industry, applied research is rapidly increasing our understanding of what quantum computers can and cannot do. It is encouraging – and inspiring – to experimentally demonstrate that the performance of the Honeywell H1 device is very close to expected behavior. ”

Analyst notes:

1.) Future work will examine whether deeper versions of R-QAOA can outperform the KK algorithm. With more two-qubit gate fidelity and more qubits, I would expect that to be easy to achieve.

2.) Quantitative computing is still in the testing phase. Quantum systems do not currently function in a production role. It will be at least 3 to 7 years before quantum computing can perform any reliable operational supply chain operation.

3.) Entropica stated that with the small number of angles available, the R-QAOA results used a very good H1 model as well as information.

4.) The training was conducted on a classic computer rather than on the H1. That is, Entropica obtained the optimal rotation parameters by simulation, and then ran these H1-based circuits. This was necessary due to limited access time to the H1. It would be useful to do the training on the QPU as well as end-to-end computer loyalty verification.

5.) The Honeywell roadmap asks the Model H1 to scale to 40 square. Since Ytterbium ions are natural qubits and even higher facts are expected, more qubits should have even more impressive results.

Disclosure: Moor Insights & Strategy, like all research and analysis firms, provides or provides paid analysis, analysis, consulting, or consulting to many high-tech companies in the industry, including Honeywell . The author has no investment positions with any of the companies that may be mentioned in this article.