The last two years have taught the industry to expect the unexpected -- a blockage in the Suez Canal disrupted maritime shipping lanes, the Coronavirus disease (COVID-19) completely changed consumer behavior, chip shortages in China rippled throughout multiple industries. And, that’s beyond the “usual” collection of tornadoes, droughts, regional conflicts, and other unfortunate events.
At the same time, logistics and supply chain executives are facing a new set of constraints. It is no longer enough to optimize for delivery times, inventory levels or carrying costs. One also wants to minimize the carbon footprint, supplier diversity and numerous other goals that might not have been at the forefront a decade or two ago.
Age-old optimization problems
The mathematical formulation of logistics and supply chain problems is old and well-understood. For instance, the Traveling Salesperson Problem, formulated in the 1800s seeks to find the optimal sequence of customer visits that a salesperson wants to make. Optimal, of course, is in the eye of the beholder. Do we need to minimize the time? The fuel consumption? The cost of travel? This problem can be extended to dropping off passengers in a rideshare application, the delivery sequence of a UPS truck and numerous other applications.
A different class of problems is often addressed through Monte Carlo simulations. Facing unpredictable supply chain conditions, companies might strive to carry enough inventory to satisfy demand, while minimizing carrying cost or the risk of inventory obsolescence. A mathematical model can be created, such as assuming the weekly arrival of inventory but with varying week-to-week quantities. Given a requested delivery schedule, that simulation is repeated multiple times to assess the likelihood of stockouts or excess inventory.
While the models are well-understood, the computational resources required to analyze them are enormous and are growing. There are additional variables to consider, additional goals to optimize for, additional suppliers to choose from. The problem is not just the cost of calculations, but the time they take. If it takes three days to find the optimal shipping schedule, this optimal solution might be worthless by the time it is produced because conditions on the ground have changed. Supply chain and logistics executives require faster solutions to these age-old problems.
Quantum computers can help
Quantum computers, a new type of computer, might offer a ray of hope. How are they different than current “classical” computers? Classical computers use binary bits that take the value of either 0 or 1 at any given time. Quantum computers use quantum bits, called qubits, that can be a simultaneous combination of both 0 and 1. This leads to one significant quantum advantage: the ability to analyze multiple scenarios simultaneously. A 10-bit classical computer can hold one of 1,024 values and analyze one such value at a time. A 10-qubit quantum computer can simultaneously hold a combination of 1,024 values and simultaneously analyze these 1,024 potential values. A 20-qubit quantum computer can analyze a million options at the same time. A 300-qubit computer, expected to be available in a couple of years, can examine more options than there are atoms in the universe. Quantum computers thus offer the possibility of a dramatic speed-up in processing times.
The implications of such acceleration are profound. Companies will be able to complete the execution of their optimization models much faster than today. As a result, they can respond faster to changing market, road, customer demand or weather conditions and make smarter decisions faster. Indeed, the Boston Consulting Group recently published a report noting that the potential value created by quantum computing in supply chain applications could be between $50-$100 billion.
Help is on the way, but it’s not here yet
But, don’t decommission your classical computer just yet. There are several issues that still need to be addressed for quantum computers to go mainstream. That “300-qubit” computer mentioned earlier is not yet available. Algorithms need to be adapted to the special ways in which quantum computers work. Current hardware is “noisy” meaning that it cannot conduct lengthy calculations without data corruption. Quantum machines are fairly scarce and mostly available on a limited basis through cloud providers.
How can you prepare for the quantum revolution?
If you think the potential benefits of quantum are too attractive to ignore, you are probably right. Imagine the angst if a competitor masters quantum computing and gains a true cost advantage.
To mitigate that risk, companies in various industries set up small quantum teams. They discover problems that might have excellent quantum solutions. They perform small-scale pilot projects to gain expertise and confidence in quantum. They use the results of these pilots to extrapolate what they could do in a year or two with the expected advances in quantum hardware and software. Some companies have announced quantum projects and are exploring the best way to harness this new technology.
Nearly 10-15 years ago, quantum computing sounded like science fiction. It is not fiction anymore. Small-scale quantum computers are available for testing and experimentation, and major vendors have unveiled near-term quantum roadmaps that can deliver substantial enterprise value soon. It’s time to dip your toe in the quantum waters.
