Could Algorithms Help Us Make Better Choices in Uncertain Times?

| March 8, 2024

Johannes Royset joins USC Viterbi, bringing expertise in operations and optimization and a background in naval applications.

Professor in the Daniel J. Epstein Department of Industrial and Systems EngineeringJohannes Royset. Image/Angel Ahabue

Professor of Industrial and Systems Engineering Johannes Royset. Image/Angel Ahabue

Whether it’s a high-stakes business scenario or a military conflict, how do you make a successful decision when faced with unknowns or adversity? For Johannes Royset, algorithms are vital for making the best choices in uncertain environments.

The optimization expert is a new professor in the Daniel J. Epstein Department of Industrial and Systems Engineering, with an extensive background in harnessing mathematical tools and data models for decision-making, particularly in naval applications.

“There are so many challenges, both small and big in society. Obviously, we need to make better and more optimal decisions,” Royset said. “The thing about optimization that I’m very interested in is that it’s a very general mathematical tool. Something that you can apply in one area one day, and then apply in a completely different area the next.”

Originally from Norway, Royset comes to USC Viterbi School of Engineering after seven years as a professor in the operations research department of the Naval Postgraduate School. Here, his research included a $2 million DARPA project developing mathematical models for an ultra high-speed naval vessel.

Royset was also recently appointed as a senior research fellow at USC’s Center for Risk and Economic Analysis of Threats and Emergencies (CREATE), where he will apply his experience at the Department of Homeland Security’s first Center of Excellence.

One of Royset’s first projects at USC Viterbi School of Engineering will be a collaboration with Kellner Family Early Career Chair and Associate Professor of Industrial and Systems Engineering John Carlsson, who has recently been researching delivery robots and supply systems in warehouses. This new research will focus on ensuring the supply of materials and food to troops during active conflicts. The work is sponsored by the Office of Naval Research and will see Royset and Carlsson develop an optimization model to enable the U.S. Navy and the Marine Corps to distribute and receive their supplies in a timely fashion.

“It’s a little bit like the problem that Amazon has, where people are requesting things, and these need to be shipped out to them from the various warehouses. The Navy has the same challenge. They have their own warehouses, and they need to ship out to their customers,” Royset said. “The additional problem that they have is that the Navy might face an adversary that is trying to disrupt these supplies, and that just adds uncertainty on top of all the other uncertainties one would face in a situation like that.”

For instance, if a military unit requests supplies in a zone of active conflict, perhaps there may not be usable roads to deliver the goods, or there may be enemy agents trying to attack the supply run. The team’s algorithms would need to consider these factors to determine whether deliveries can be made successfully and safely using various methods, including autonomous systems, such as aerial drones.

“We would need to ensure nobody is in harm’s way if somebody might try to disrupt the supply. If autonomous systems could do it, we must consider what type of capacity they have to move it, and can they move quickly and reliably enough,” Royset said. “This would then involve various types of mathematical optimization techniques and algorithms that we’re developing. We are looking through thousands of alternative decisions and trying to pick out the ones that are best,” Royset said.

“Digital twins” can make decisions simpler

Another research area Royset is looking forward to exploring is the “digital twin” — a sophisticated technique that creates a detailed digital model of a physical entity, for instance, an aircraft or a vehicle, allowing a multitude of scenarios for experimentation.

“Imagine if you had a digital twin for your car that could tell you when you should take it in,” Royset said. “Maybe you don’t have to do that for months because you’re driving on clean roads and no dust gets into the air filter. Or maybe you’re doing off-roading, so there’s much more wear and tear on your car. The digital twin of your car can tell you about that and allow you to customize your maintenance.”

Royset also brings his expertise as a lecturer to the Epstein Department, where he will teach optimization techniques based on his textbook, An Optimization Primer.

“Optimization technology has become such an important thing for all engineers to know about because it underpins machine learning in a fundamental way,” Royset said. “Every time you’re using Chat GPT or any other AI-based tool, all those have typically been developed using some sort of optimization methodology.”

Royset completed his Ph.D. at the University of California at Berkeley in 2002. Throughout his career, he has received many accolades, including a National Research Council postdoctoral fellowship, a Young Investigator Award from the Air Force Office of Scientific Research, the Barchi Prize, and the MOR Journal Award from the Military Operations Research Society.

Royset said that one of the highlights of his work is the scope it offers to creatively approach problems with a curious mindset.

“The contribution that I really enjoy is that we in academia have the luxury of simply thinking deeply about something. To really explore new ideas,” Royset said. “Then, hopefully, we can come up with a completely new way of looking at the problem that may be a real game changer.”

Published on March 8th, 2024

Last updated on March 8th, 2024

Share this Story