As it looks for an edge in future warfare, Army has identified an initial three challenges it needs to address and is turning to quantum technology for solutions.
In addition to a means of detecting subterranean structures and the movement of humans and material through them and the ability to disrupt an adversary’s space-based communications systems, Army is seeking to optimise the future resupply of soldiers in battle.
Coincident with Army Innovation Day and Army Robotics Exposition in Brisbane in April, the inaugural Army Quantum Technology Challenge 2021 (QTC 2021) not only set out the initial three challenges for academia and industry, but also launched Army’s Quantum Technology Roadmap for the future.
“Quantum technologies are part of an emerging group of disruptive technologies that have the potential to drive significant change in the character of warfare,” Head of Land Capability, Major General Simon Stuart explained at the launch. “We recognise that we need to act now in order to be an informed and a demanding customer.”
Addressing the challenge
The logistics challenge is considering how to most effectively plan the resupply of deployed units over the ‘last mile’. Three teams presented solutions at QTC 2021, including NEC Corporation subsidiary D-Wave, the University of Melbourne’s Heuristic Algorithm Quantum Computing (HAQC) Team, and Sydney-based SME Q-CTRL.
At the event D-Wave demonstrated the use of hybrid quantum computing based upon its latest Advance computer, which customers can access via the company’s Leap quantum-based cloud service. The HAQC Team is developing a combined approach which exploits the combined power of quantum computers and high-performance classical computers (HPCs), through a collaborative program with IBM’s Quantum Network. Q-CTRL’s demonstration at QTC 2021 looked at the resupply of multiple distributed force elements from a central logistics hub, utilising unmanned ground vehicles (UGVs).
The quantum advantage
Quantum control is defined as the control of physical systems whose behaviour is in turn defined by the laws of quantum mechanics and quantum computing. The use of quantum physics to encode and process information to solve problems that are otherwise extremely difficult, or even impossible, to solve using conventional computers promises to deliver a tangible advantage to Army.
“Army is specifically interested in the applicability of quantum computing to the kind of problems that are relevant to them, in this case the logistics resupply problem, and we know that quantum computers can provide some very serious boosts in performance,” Q-CTRL founder and CEO Michael Biercuk explained.
“We don’t look at quantum computers as general-purpose computers, but as special-purpose machines that overcome real computational bottlenecks. We approached this problem understanding very broadly that the frame of what Army was interested in for a Quantum Technology Challenge was compatible with one of the known algorithmic approaches for executing programs in quantum computers.”
Solving the problem
One of the advantages of quantum computers is their ability to solve optimisation problems, in this instance typified by the requirement to find the best pathway for UGVs, when there are perhaps multiple vehicles and multiple pathways to choose from, to reach their destination.
“Our approach was not to try to develop a new or better algorithm for executing Army’s problem on a quantum computer, but instead leverage all of our capabilities in quantum control to get the most from real hardware,” Biercuk said.
Biercuk says Q-CTRL took a three-staged approach to the problem, the first of which was to efficiently map the problem into a known quantum algorithm. “That can give you some big gains and we demonstrated some large performance benefits by doing that efficiently, limiting the number of hardware elements needed to represent a problem that matters to Army,” he added.
The second step was to calculate a small part of the problem and then embed it into a much larger quantum classical optimisation loop, to select the best model; and the third step was to develop drop-in replacements for the quantum logic operations that are executed on a quantum computer.
Optimising resupply options
Biercuk said the results of the exercise demonstrated an algorithmic success of around 80-times using the combination of the three approaches, and the efficient encoding of the problem achieved an eight-times improvement in run clock time against conventional methodologies.
“This was done on real quantum computing hardware and we supplemented it by lots of simulations of larger problems, but this was an early demonstration and, by using algorithms that already exist and attacking all the practical elements that make the algorithm fail or succeed, can give order of magnitude benefits,” Biercuk said.
“This is a known problem with a known algorithm. When we made the three stages of changes, investigating the classical optimiser, the compilation strategy (and) the hardware implementation of quantum logic, the success probability went from 1.3 per cent to 88 per cent –that’s an 80-times gain using the combined approach.”
Next steps
While the problems addressed in the initial demonstration were relatively small, it is hoped that wider Army logistics problems can be solved in the future, particularly in the light of hardware roadmaps now being set by large manufacturers such as IBM and Google.
“Quantum computers at present are in the prototype stage and don’t give useful performance. They’re too small and fragile to give useful solutions, and that’s a threshold we call ‘Quantum Advantage’, when you would actually use a quantum computer instead of the best supercomputing solution, (but) we’re not there yet,” Biercuk said.
Quantum computer system size is measured in qubits (quantum bits), which are the fundamental carriers of quantum information. The computers of today have, on average, 50 qubits, but a size of around 400 qubits is thought to be the threshold for the aforementioned quantum advantage. Under the industry roadmaps, system size looks likely to increase to 1000-cubit machines within the next five years.
“What we’re ultimately aiming for is a solution that allows anyone in Defence to gain access to useful quantum computing for the problems they care about. We aim to build this infrastructure software that delivers the results we’ve demonstrated to the end user, so they can use it themselves,” Biercuk concluded.