Author: Ed Klotz, PhD
Supply chain disruption has been making a lot of headlines over the past year. The COVID-19 pandemic, the Suez Canal blockage, and other recent events have shone a spotlight on the impact of supply chain disruption and the importance of supply chain agility and resilience.
But supply chain disruption is, of course, nothing new – in fact, it always has been (and always will be) part and parcel of each and every supply chain. Supply chain leaders have come to expect the unexpected, and know how to leverage the most effective strategies and technologies to react and respond to disruptions in the most efficient manner possible.
The key technology that supply chain stakeholders have relied on over the years to handle disruption (and also to manage their end-to-end supply chain planning, decision making, and operations) is mathematical optimization, a powerful prescriptive analytics tool.
Ever since the late 1980s – when I started my career in the mathematical optimization software industry – companies have been utilizing this AI technology to conquer their supply chain challenges and combat supply disruption (and to solve a whole host of other business problems as well).
Mathematical optimization solvers (which are algorithm-based problem-solving engines) are embedded in a wide array of supply chain planning software applications and used by enterprises across industries to optimize their end-to-end supply chains.
Interestingly, many business users – who leverage mathematical optimization in various off-the-shelf or custom-built software solutions – are not even aware that mathematical optimization is the engine that makes their company’s supply chain planning systems run. Just like you can drive a car without understanding what’s going on under the hood, you can use a mathematical optimization application – to make optimal supply chain plans and decisions – without having a deep understanding of its inner workings.
Whether you know it or not, the fact is that mathematical optimization has established itself as an indispensable tool for handling supply chain planning in general and supply chain disruptions in particular.
The question is: Why does mathematical optimization – which was first introduced over 70 years ago – have such staying power as the standard software technology for managing supply chain operations and disruptions?
Increasing Business Complexity, Improving Software Technology
Over the years, mathematical optimization technologies have constantly evolved, and we have seen consistent, colossal improvements in the speed and robustness of mathematical optimization solvers through the decades. To give you an example that illustrates this: Some business problems that can be solved in a single second using today’s cutting-edge solver technologies would have taken 55 years to solve in 1991.
What’s been propelling this phenomenal technological development? There are several key factors:
- The hard work and innovation of mathematical optimization software developers, who are always looking to push the envelope in terms of solver speed and functionality.
- Advancements in computing power and data availability and quality, which open the door for additional improvements to the software technology.
- The ever-increasing complexity of real-world business problems, which has inspired developers to enhance mathematical optimization technologies to ensure they are able to tackle the problems of the day.
One business area where we’ve witnessed an explosion in complexity is the supply chain domain – with sprawling global production and distribution networks, volatile supply and demand dynamics, and incessant supply chain disruptions.
But over the years, as supply chain complexity and disruptions have increased, mathematical optimization technologies have evolved and improved in parallel – and, consequently, these technologies have been able to “meet the moment” and solve the supply chain problems that companies are facing at any given time.
When supply chain disruptions hit, businesses always know they can depend on mathematical optimization to give them the answers they need, when they need them – so that they can sense and react to supply chain issues in real-time, and take the necessary actions to resolve disruptions as quickly and effectively as possible.
Designed to Deal with Disruption
To really gain an understanding of why mathematical optimization is such a powerful weapon in combating supply chain disruption, we need to take a look inside a mathematical optimization application.
Mathematical optimization has a few fundamental features that make it uniquely well suited to handle supply chain disruptions:
1. The mathematical optimization model: Each mathematical optimization supply chain application is built on a model (or, in other words, a digital twin) that encompasses your end-to-end network or particular components of your supply chain (such as your manufacturing, warehousing, or distribution operations).When supply chain disruptions occur, your mathematical optimization application provides you with an end-to-end view and real-time control over your entire supply chain network – so that you can:
- Instantly identify the root causes of disruption and bottlenecks (using “infeasibility analysis”) and make the best decisions on how to address them.
- Easily adjust your mathematical optimization model so that it mirrors current conditions and disruptions across your supply chain.
- Automatically reoptimize plans and schedules and reallocate resources so that you can minimize time to recovery, maximize operational efficiency, and foster ongoing business continuity.
2. The latest available data: By definition, disruptions are outliers – and this means that you may face challenges obtaining historical or training set data (which is used in AI technologies such as machine learning) that you can utilize to generate solutions and guide your decisions in times of disruption.In contrast, mathematical optimization leverages the latest available data flowing into the application from various sources (such as ERP and MES systems and IoT devices) across your supply chain network. So, when you are experiencing disruptions, you can have confidence that your mathematical application will be able to deliver optimal solutions and drive optimal decision making even in the face of these unprecedented supply chain challenges and changes.
3. The ability to measure solution quality: When a disruption occurs, you have to respond rapidly as time is limited – and it’s sometimes not possible to completely reoptimize your model. In these instances, mathematical optimization (unlike heuristic techniques) provides a measure of how close your solution is to optimality – so you know exactly how much money you’re leaving on the table.
4. What-if analysis: In order to deal with supply chain disruptions, you need to not only be able to quickly react in real-time to these events as they occur, but also to be able to proactively plan and prepare for future problems by pinpointing potential risks in your supply chain.Mathematical optimization’s scenario analysis capability allows you to accomplish the latter objective by:
- Creating and investigating multiple supply, demand, inventory, capacity, and other what-if scenarios to gauge their impact on your business.
- Discover hidden risks and make strategic decisions that boost long-term supply chain resilience and profitability.
Looking to the Future
Mathematical optimization technologies – because of the way they are designed and the way they have developed over the past decades – have cemented their place as the software tool of choice in handling supply chain disruption.
Nobody knows the future and can predict where, when, and how supply chain disruptions will strike.
What we do know for sure is that supply chain disruptions will continue to occur, supply chain complexity will continue to grow, and mathematical optimization technologies will continue be used by supply chain leaders to overcome these challenges.
We also know that (although it’s a “mature” AI technology) mathematical optimization will continue to evolve – so that it can enable enterprises to generate optimal, data-driven solutions to the problems of the day.
A version of this article was originally published on Supply and Demand Chain Executive here.