What is Optimization and Why Should I Care?
In today’s hypercompetitive business environment, companies are desperately seeking competitive advantage. And there are only three ways to get there: product differentiation, value pricing, or operational excellence. To achieve operational excellence, the key to unlocking success often lies in supply chain management – finding the best possible solution to your company’s planning and scheduling challenges. And that’s that optimization boils down to – seeking, analyzing and determining a path that will boost customer satisfaction and bottom-line results.
Most companies use what is sometimes called the heuristic approach – looking for a simple rule of thumb that provides a quick fix that is easy to understand and implement. It’s a little like smoking. The quick fix feels good, but should come with a corporate Surgeon General’s warning that they can actually be dangerous to your corporate health. Often the heuristics generated are too “myopic”, and these well-intentioned approximations can inadvertently lead to solutions far from optimality – i.e., that leave a chunk of money on the table when put into practice. Also, as your business processes evolve, they will inevitably fail because they are generally not systematic or flexible enough to accommodate your company’s growth requirements. For all their prevalence (even major consulting companies resort to them), they are like little Band-Aids, hundreds of them, all over your corporate body, hiding the root cause issue.
A significantly more effective approach is through a rigorous, mathematical optimization process that can be designed to deliver the actual best possible solution. Many sophisticated tools are available – such as constraint programming, local search, machine learning algorithms, evolutionary algorithms, like genetic algorithms, and methods based on analogies to physical processes such as simulated annealing. Higher-level heuristics, i.e., metaheuristics, can be included – especially for real-world challenges, but unlike myopic heuristics, these are grounded in mathematical representations and deeply studied computational paradigms. They may orchestrate a collection of “exact” methods and sub-heuristics to perform their operations, and can consistently outperform “ad hoc” heuristic approaches.
The advantage of formal optimization that marries mathematical algorithms with higher-level meta-heuristics is the flexibility to automatically adjust and adapt to the myriad decision variables and changing goals, constraints, and complexities. Since working at this level is expensive, and requires exceptional expertise and complex technologies, only market leading companies – the ones that are usually ahead of the curve – can afford to invest in significant Operations Research efforts. You need to hire at least a couple of expert PhDs, and this kind of talent is in short supply.
This approach, ultimately, is the key to cost minimization, revenue maximization, performance enhancement and gaining understanding through sensitivity analysis about the underlying forces and structures that drive success.
Now, one more thing… if you pay even a modicum of attention to technology news, you may have heard about Google’s recent claim of “quantum supremacy” – they’ve used a quantum computer to solve a specific problem 1.5 billion times faster than the current reigning classical supercomputer. This was sort of a Sputnik event, the firing of a metaphoric starting gun that has galvanized the industry to intensify their quantum effort.
What’s really interesting is that Google accomplished the breakthrough several years earlier than predicted by “reasonable” experts. These experts, after shaking off their failed predictions, are now saying that quantum computers powerful enough to disrupt major industries will begin to arrive in three to five years. Because it will take a few years for any company to build the expertise required to benefit from quantum computing… the time to start is now.
So what’s the big deal? Well, here’s an illustrative example of the kind of improvement that quantum computing could achieve. In 2017, UPS’ on-time failure rate for items delivered by ground was 0.9% in the week ended Dec. 23, and FedEx’s failure rate was 1.3%. For the USPS, their failure rate was closer to 3%. Since the use of quantum optimization could potentially improve logistics performance by 2 or 3 orders of magnitude, the USPS could, someday in the future, surpass its competitors by delivering Christmas gifts with a 99.97% success rate.
A word of caution. It really is quite difficult to predict when breakthroughs will happen in this field. This is because, as Jonathan Dowling, a theoretical physicist at Louisiana State University, has said, “Quantum supremacy is like a horse race where you don’t know how fast your horse is, you don’t know how fast anybody else’s horse is, and some of the horses are goats.”
So are horses or goats the only alternatives? There’s another alternative that takes the risk out of the equation: hybrid quantum computing. By blending of quantum and “quantum inspired” classical computing, productive results can be achieved now, without having to wait three to seven years to find out if you bet the ranch on a goat. At the core of most quantum optimization software is something called "QUBO" – the quadratic unconstrained binary optimization algorithm. This model maps perfectly to how certain quantum processors solve something called the Ising Hamiltonian, and provides some low-hanging fruit in the field of formal mathematical optimization.
Distinguished Professor Fred Glover of the University of Colorado at Boulder, the developer of QUBO 2.0, is championing a new approach called Quantum Bridge Analytics (QBA) that implements a hybrid approach. This will enable researchers to model unified solutions that can run on today’s classical computing systems as well as the full quantum computing systems that will arrive in a few to several years. So instead of going from 3% to 0.03% failure rate, maybe some postal operator could get to .0.3%. while waiting for quantum hardware to be realized. That would still beat FedEx’s current performance, right?
This is truly an amazing, exciting time. We are at the doorway to an as-yet barely imagined new world, and opportunity is knocking. Who will answer the call? The technology is evolving at an exponential rate and competition today is global. Will you sit on the sidelines while your more nimble competitors have all the fun and win the prizes? The Dutch have a saying: “zich de kaas niet van het brood laten eten”… don’t let your competitor eat the cheese from your bread.
Now, ask yourself these questions to assess if it’s time to get started.
- Just how complex are the problems facing your company?
- Is the performance of your business processes lagging behind that of your competitors?
- How well do your resources and strategies match your vision of aggressive growth?
- How adventurous are you feeling?
And if you decide you are ready, how do you get started? Here are some suggestions.
- How about initiating a collaborative effort to explore quantum computing for your industry?
- Conduct brainstorming sessions about the impact of quantum computing as a strategic advantage, or to envision new disruptive business models – so you can Uber the competition before they Uber you.
- Last, but not least, maybe Quantum Bridge Analytics could inspire a transformative business model for you that delivers positive ROI in the near term.