BMW’s 3,854 variables problem solved in six minutes using quantum computing

Quantum computing company Quantum Computing Inc. (QCI) made a name for itself by solving a 3,854-variable optimization problem for BMW. The company used its new hardware-based quantum computing solution, Entropy Quantum Computing (EQC), to solve the ideal placement of vehicle sensors in BMW’s 2022 Vehicle Sensor Placement Challenge (VSPC). Its new quantum system delivered 70 times the performance of its 2021 entry, which leveraged the company’s hybrid quantum implementation of spin-off quantum computing player D-Wave.

“We are very proud to have achieved what we believe to be a milestone in the evolution of quantum,” said Bob Liscouski, CEO of QCI in a press release. “We believe this proves that innovative quantum computing technologies can solve real business problems today. What is even more significant is the complexity of the problem solved. It was not just a rudimentary problem to show that quantum solutions will be feasible one day; this was a very real and important problem whose solution can potentially help accelerate the realization of the autonomous vehicle industry.

This marks a use of quantum computing to solve real-world, actionable problems that would take exponentially longer times to be solved by classical computers. According to QCI, this proves the advantages of its approach to quantum computing over other quantum systems available today. Alternatives such as IBM’s Eagle 127 qubit quantum processing unit (QPU) and Quantum Brilliance’s diamond-based QPUs (already deployed in data center environments) are all classified as Noisy Intermediate Scale Quantum systems. (NISQ). QCI claims its demonstration is proof that it is achieving a quantum advantage (the moment when quantum computers solve problems that would be impossible for classical systems).

Placing sensors in vehicles – and especially autonomous vehicles – is an incredible challenge. A multitude of variables must be considered – variables such as chassis design (which has vehicle safety implications), freedom from obstruction (different locations provide different fields of view or allow for less possibility of error), wind resistance and weight balance, just to name a few.

This is a problem that requires many trial and error processes that may not provide the optimal solution and must be redone for each new vehicle and each new sensor advancement. This is part of the reason why vehicle design has remained relatively lifeless for years – deviating from already known solutions increases costs, which then reduces profits.

Due to the number of variables and constraints (QCI cites 3,854 variables and 500 constraints placed on the solution), calculating all possible positions for sensor placement on a typical system is hitting the walls in terms of cost – time computing is an expensive pursuit, as F1 teams will tell you.

Even before the money was counted, the very real cost of computing time in traditional systems made many problems unsolvable (such as logistics management, sequencing of steps and prioritization).

These are problems that quantum computing, with its probabilistic approach to computing, can solve in a fraction of the time. So much so that QCI solved BMW’s optimization problem in less than six minutes, providing the best possible solution to the placement problem at hand. In doing so, it delivered a solution comprised of 15 sensors, which enabled 96% vehicle coverage by leveraging QCI’s quantum hardware and software system.

In response to VSPC, QCI leveraged a new hardware form of quantum computing. Entropy Quantum Computing, as it is called, eliminates the requirements for a near-perfect environment in which qubits operate, dramatically reducing design, installation, and operating costs. Entropy refers to the natural evolution of any system, which tends to head towards chaos (or in this case, disorder).

When you can get away with a noisier environment (in which temperatures, electromagnetic radiation, and other variables better forgive the coherence of the quantum system), deploying quantum computers becomes much more feasible.

Consistency is a fundamental requirement of quantum computers, as changes in its environment can cause it to inadvertently change state – introducing costly and sometimes fatal errors into the calculations.

QCI’s Entropy Quantum Computing approach works by factoring the environment itself into the computational results. Time and money are saved by not having to control all the variables outside of the quantum processing unit itself – instead the system adapts to the changing environment, analyzing his comments and what this means for the quantum states of qubits.

To greatly simplify things, think about how modern processors dynamically change voltages and frequency based on workload, while taking into account variables such as power consumption and operating temperature.

The commercial and general feasibility of QCI’s quantum computing solution remains to be seen; Interestingly, companies with more resources and history than QCI have taken other approaches to quantum computing. Others, like Microsoft, are still pursuing their own specific qubits. Each of them praises the merits of the approach he has chosen.

It’s not so much a race (although there is a race for funding and market share) as it is exploring different venues for quantum computing. It is perhaps a testament to its complexity that there are so many possible approaches to exploiting what is likely to become the next big frontier for computer science.

Sherry J. Basler