Researchers have developed an artificial intelligence system that manages complex networks where decision-makers operate at different authority levels, addressing limitations in current AI approaches that assume equal decision-making power.
The framework, published in IEEE Transactions on Systems, Man and Cybernetics: Systems, targets applications including smart energy grids, traffic networks and autonomous vehicle coordination. These systems involve hierarchical decision-making where leaders act first and followers respond accordingly.
Associate professors Zhen Ni and Xiangnan Zhong from Florida Atlantic University’s College of Engineering and Computer Science designed the system using reinforcement learning combined with game theory principles. Their approach incorporates the Stackelberg-Nash game model, establishing clear leader-follower relationships between decision-makers.
“Traditional AI methods often treat every decision-maker as equal, operating at the same time with the same level of influence,” explained Ni. “While this makes for clean simulations, it doesn’t reflect how decisions are actually made in real-world scenarios – especially in environments full of uncertainty, limited bandwidth and uneven access to information.”
The framework introduces an event-triggered mechanism reducing computational demands by updating decisions only when necessary rather than continuously. This approach addresses scenarios where different participants possess varying information levels and predictability constraints.
“Instead of constantly updating decisions at every time step, which is typical of many AI systems, our method updates decisions only when necessary, saving energy and processing power while maintaining performance and stability,” said Zhong.
The system handles power asymmetries between decision-makers whilst managing mismatched uncertainties where participants operate with different information access levels. Smart grid applications might involve utility companies making primary decisions about power distribution whilst households adjust consumption patterns in response.
College Dean Stella Batalama noted the research addresses practical infrastructure challenges: “By developing a method that reflects real-world decision hierarchies and adapts to imperfect information, Professors Zhong and Ni are helping us move closer to practical, intelligent systems that can handle the complexity of our modern infrastructure.”
The research team validated their approach through theoretical analysis and simulation studies, demonstrating maintained system stability and optimal strategy outcomes. Support came from the National Science Foundation and US Department of Transportation.
Researchers plan to expand testing for larger-scale real-world applications, aiming to integrate the framework into operational urban systems managing cities, traffic coordination and autonomous machine fleets.
