Richelle L Smith


“Polychronous Oscillatory Cellular Neural Networks for Solving Graph Coloring Problems”

Energy demands for computation are extrapolated to exceed global energy production by 2040, while emerging applications such as Internet of Things, machine learning, and blockchain will further increase the computing energy demand. Together, these trends underscore the need for computers with improved energy efficiency and computing capability. Quantum and cryogenic computing are potential alternatives to traditional von Neumann computing, but they require elaborate cooling systems, limiting their form factor, portability and widespread adoption. Addressing these issues, this invention is a fast and energy-efficient computational platform suitable for solving large-scale, combinatorial optimization problems. The core technology is a computing system comprised of an array of superharmonic injection-locked, coupled oscillators that compute solutions to combinatorial optimization problems in real-time. The array of oscillators is programmable and reconfigurable, allowing a wide variety of problems to be mapped to it. As an illustrative example, this work focuses on using our computational platform to solve graph coloring problems.


Richelle’s research investigates the design of energy-efficient hardware systems, with a focus on circuits and architectures for low-power computing and communication. By lowering the power consumption of the computer chips that handle our data traffic, we can reduce our carbon footprint on the planet.

6 + 3 =