Operations Research
In the Slipstream of AI
Everyone knows artificial intelligence today. Operations Reserach (OR) is only known to specialists. One of them is Hans-Jürgen Zimmermann. When he founded INFORM in 1969, he succeeded early in what today is the great promise of AI - using algorithms to make processes faster and more reliable.
Technologies based on artificial intelligence (AI) are considered the epitome of modern progress. AI fascinates people and makes a regular appearance in news headlines. In the business environment, methods such as machine learning, neural networks, or big data are common knowledge. In contrast, the time-honored mathematical discipline of operations research (OR) seems rather inconspicuous
______
"Operations research means seeing the big picture. Our programs go beyond local optimization to global."
Artificial Intelligence and OR are related
OR does not make the headlines. No one gets excited about it. "The term isn't very catchy, is it?" says Hans-Jürgen Zimmermann, an economist and OR pioneer. "Little has changed about that in the last 60 years." Yet AI and OR are related and complementary. Without OR, companies today would hardly be able to make informed decisions in an increasingly complex world with numerous interdependencies. And without machine learning and big data, the required data would not be available in the required quality. OR is an intelligent decision support tool for people managing complex systems. OR is also interdisciplinary; users set a goal, whether it's higher product quality, better service, lower spending, or less risk. To do this, they identify all the impact factors - economic, social, psychological, or physical. Using mathematical techniques, the software ultimately calculates an optimized solution. While AI calculates predictions based on logic to automate processes, OR models runs processes and optimizes them. Today, the two approaches are intermixing more and more.
Broad range of applications
This combination ensures that the range of applications in which OR can and will be used is very broad, now and in the future. In production, logistics, public transport, health care, energy, education, and politics, OR applications shape more aspects of today’s life than many realize. Whether modeling balanced electoral districts, organizing reliable patient appointments for hospitals, calculating efficient production sequences for manufacturing companies, or planning fast delivery routes for freight forwarders, it is hard to imagine all these everyday situations without OR and its mathematical models and algorithms.
While today’s readily available data is used to obtain a better understanding of business processes, there is also a growing awareness that mathematics and algorithms are indispensable for making the best possible decisions based on this knowledge. Even though the term "OR" is not at the forefront of everyday discussions, OR has become indispensable in the slipstream of AI.
Pioneering
However, getting to this point took a long time. The mathematical methods used in OR had to first be made applicable for industrial purposes. INFORM, and at the direction of its founder Hans-Jürgen Zimmermann, was one of the first companies to succeed. "I wanted to prove that OR helps make better decisions," says Zimmermann. “Companies that use it can save a lot of money.” In 1969, he founded INFORM with a small team.
In the beginning, Zimmermann found every problem intriguing, such as how a winery could use OR to produce the optimal sparkling wine. "Sparkling wine is usually a blend of wines. The winery pours different wines together, so the sparkling wine has certain characteristics," Zimmermann says. "Usually, the cellar masters rely on experience. They blend, taste, adjust the blend, taste again, and so on." But there is a downside to this method: It takes a lot of time, and the constant mixing, tasting, and pouring consume a lot of raw materials, for which appropriate storage capacity must also be available. INFORM took a mathematical approach: a linear target function based on the characteristics of the wines, such as sugar and acid content, was formulated. The winery could specify cost reduction, taste optimization, or both as the target condition. en.
______
Decisions based on OR often contradict gut feelings. What does not seem efficient for the individual can be a great advantage to the whole.
After skepticism came economic success
However, economic success could not be achieved with such projects. Many companies Zimmermann approached were interested, but also skeptical. "Decisions based on OR often contradict gut feelings," Zimmermann says. "What does not seem efficient to the individual can be a huge advantage to the company." Zimmermann identified one such advantage in the way one company he visited handled its own fleet of vehicles. Each department assigned individual driving jobs. One driver drove to the neighboring town for an installation and returned with a full load. Sales sent a second driver to the same city a few hours later with a full load. He than returned with an empty car. That's hard for an OR'er to watch," Zimmermann says. INFORM developed models that optimized the fleet. The software collected driving jobs and combined them intelligently. It seemed inefficient to each department that their order had to wait, but the overall system profited. The company saved on superfluous driving. "Departments often think in silos," Zimmermann says. "OR means looking at the big picture. Our programs go from local optimization to global.”
To this day, INFORM is characterized by its passion for removing the smallest grain of sand out of any gear, no matter how large, because OR has made itself indispensable in the slipstream of AI!
First standard software
At the end of the 1970s, INFORM developed one of its first standard products within the context of internal transport optimization. Today, the SyncroTESS software optimizes logistical processes in transport logistics for numerous industries - from construction logistics and container handling to yard management for large automobile manufacturers. In 1991, the software was also used for the first time at Frankfurt Airport to manage the complex ground handling of aircraft. Entry into the aviation industry also marked the beginning of INFORM becoming a global player. Today, under the name GroundStar, the INFORM system manages up to 20,000 flights a day at more than 170 airports worldwide. In the years that followed, INFORM has become active in new markets and developed new systems that are used in industry, trade, ports, banks, and insurance companies.
It helped that conditions for OR began to improve in the mid-1980s. Until then, data had been scarce. Part of the job was first to capture data digitally. Today, data is created every second. Manufacturing, commerce, universities, libraries, and hospitals all produce data incessantly. The good thing about this is that the more data available, the better processes like OR work, and the broader the range of applications.
Proof provided
INFORM and Zimmermann have long since succeeded in proving what OR is capable of. To this day, INFORM is characterized by its passion for removing the smallest grain of sand out of any gear, no matter how large. And there will continue to be plenty of opportunities to do so in the future. The combination of big data, statistics, and machine learning with OR optimization algorithms offers new application possibilities and supports companies in their digital transformation.
Operations Reserach: The History
Many business processes can be traced back to fundamental mathematical problems, which were and are solved with operations research. Mathematicians have been working on these problems longer than the existence of the OR discipline.
Read more about the history of OR and the problems it has solved.