The AATLAS project, funded by the ANR, focuses on developing a hybrid analog-digital computational platform to tackle NP-hard optimization problems, such as Max-Cut, 3-SAT, and TSP. Leveraging innovative field-programmable analog arrays (FPAAs) and scalable Hopfield networks and Ising machines, the project aims to deliver energy-efficient solutions beyond the capabilities of traditional digital systems. By combining analog and digital technologies, AATLAS seeks to enable the resolution of large-scale problems with improved energy efficiency and scalability, opening new pathways in computation for fields such as logistics, manufacturing, and data processing.
The project brings together the expertise of three partners:
- Georgia Tech/ Georgia Tech-CNRS IRL 2958 (GT-CNRS): Specializing in FPAA design and nonlinear dynamics.
- CentraleSupélec (CS): Focused on the design and analysis of complex systems.
- Institut Jean Lamour (IJL): Providing expertise in reconfigurable computing architectures and networks of artificial spins.