Yale Engineering researchers have been awarded a $9 million Multi-University Research Initiative (MURI) grant from the Office of Naval Research to explore how complex, high-dimensional interventions can control or design coherent physical systems.
Led by Hui Cao, the John C. Malone Professor of Applied Physics, the project aims to understand a counterintuitive phenomenon observed in many complex systems: that introducing disorder or complexity can often lead to more orderly behavior.
This research initiative is inspired by the widespread, but still poorly understood, observation that many complex systems can be dramatically changed by introducing complex “disordered” perturbations. The team’s critical insight is that emerging machine learning techniques can now provide an effective way to design these complex, seemingly disorderly interventions, even in systems where human intuition or traditional physics analysis fails.
As a particular challenge, the team aims to create laser systems in which many distinct lasers are synchronized into the same phase, allowing them to act effectively as one single laser. This could enable unprecedentedly high-power, controllable laser systems with potential applications in plasma or fusion research.
The broader vision of the project is to develop computational, AI-driven techniques for designing and controlling high-dimensional, disordered nonlinear dynamical systems. These range from multimode lasers to networks of superconducting oscillators and radio-frequency electronic devices. Potential applications include directed energy for powering remote systems or controlling plasmas, and adaptive remote sensing.
The long-term aim, synthesizing both specific experimental prototypes and broad theoretical advances, is to provide the data resources and algorithmic bases for an AI-driven, universal algorithm to control complex, nonlinear dynamical systems with minimal, but still high-dimensional interventions.
The success of this research could revolutionize engineering across various domains, including computer and network design, robotics, and enable entirely new approaches to automated scientific experiments in systems that are presently difficult to systematically control, such as complex lasers, and chaotic or turbulent electronic networks and fluids.
The multidisciplinary team includes researchers from several institutions, including Herbert Winful of the University of Michigan, Steven Anlage of the University of Maryland, Vassilios Kovanis of Virginia Tech, Tsampikos Kottos of Wesleyan University, Ying-Cheng Lai of Arizona State University, and Logan Wright of Yale’s Department of Applied Physics.
This story is taken from the Yale Engineering news story of August 9, 2024.