����������

How can online feedback optimization inform traffic flows for the 2028 Los Angeles Summer Olympics?��

Associate Professor Emiliano Dall’Anese and his research group examined that concept for a paper that recently won the prestigious ‘Best Paper Award’ in the IEEE journal .��

The paper, “,” coauthored with Gianluca Bianchin, Jorge Cortés and Jorge Poveda, was selected for its originality, potential impact on the foundations of network systems and practical significance in applications.��

“I'm very thankful to our former postdoc Gianluca and our collaborators for the time and efforts put in developing the framework presented in this paper. More broadly, I am thankful to our PhD students and collaborators that have contributed in developing theory and tools in the areas of online feedback optimization and data-driven optimization throughout the years,” said Dall’Anese. “It was definitely a team achievement.”

Dall’Anese’s research focuses on the intersection of optimization, learning and control in complex network systems. Current application domains for his research include power and energy systems and cyber-physical systems.��

Applications Across Areas��

“Our paper contributes positively for new tools and methods in the context of controls for complex and autonomous systems,” said Dall’Anese, “and if I were to go one step further, contributes in AI for infrastructures and cyber-physical systems.”��

The mathematical framework behind this paper is centered on online feedback optimization — a topic his research group pioneered over the past 10 years. In this paper, the research group examined the Los Angeles highway system and used their framework to model traffic in hopes of lessening congestion ahead of the city’s 2028 Summer Olympics. Their outputs showed that during some parts of the day, the tools they developed significantly outperformed existing techniques.��

Online feedback optimization has contributed tools in other application areas.��

“Of course, the math needs to be customized, but it’s also applicable to problems in power systems, robotic systems and control of epidemics,” said Dall’Anese. “We have shown how to apply our tools in these areas such as power systems and autonomous systems in other publications.”������

A Personal Accomplishment����

Upon learning of this award-winning paper, the first thing Dall’Anese did was call his parents back in Italy, letting them know all their sacrifices they made in the past were being rewarded right now.��

“My family struggled financially for many years,” he said. “I'm always thankful to the support that my parents provided through the years for myself and my sister.”

Some words of wisdom Dall’Anese hope to instill from this accomplishment, especially being a first-generation college student is, “anybody can be sitting here at my desk with some perseverance, hard-working and the valuable guidance of their research advisors. Anybody can make it.”����

What Lies Ahead

His research group has been working on extensions of the paper’s mathematical concepts to systems in which safety needs to be prioritized�� — such as power grids or autonomous vehicles.

“You don’t want these vehicles to cross highway lanes. This brings together several core areas, namely optimization, machine learning and engineering,” said Dall’Anese.��

In power systems, ensuring grids are tightly regulated within a given operating region is critical. Otherwise, cascading failures or disrupted service could occur without proper safety controls.

Prior to joining the University of Colorado Boulder, Dall’Anese was a senior scientist at the National Renewable Energy Laboratory creating an impact in the world of sustainable power energy systems. His latest paper is coming full circle with his research motivation.��

“When looking at power and energy systems, our work’s grand goal right now is to resolve climate change issues and enable sustainability and resilience in our current power infrastructure,” Dall’Anese said.��

Photo:��Emiliano Dall’Anese (Credit: Jesse Peterson)��

���������� the IEEE Transactions on Control of Network Systems

The IEEE Transactions on Control of Network Systems publishes peer-reviewed papers at the intersection of control systems and network science. Topics covered by this journal include collaborative control, distributed learning, multi-agent systems, distributed optimization, control of collective behavior, large-scale complex systems and control with communication constraints.��