SEOUL, May 06 (AJP) - Researchers at the Korea Advanced Institute of Science and Technology have developed a silicon-based hardware platform designed to solve complex combinatorial optimization problems, the institute said Wednesday.
A joint research team led by the Korea Advanced Institute of Science and Technology (KAIST) Professors Choi Yang-kyu and Kim Sang-hyun from the School of Electrical Engineering succeeded in implementing an oscillatory Ising machine using standard silicon semiconductor processes. The system uses multiple vibrating elements, or oscillators, that interact with each other to naturally reach an energy-stable state representing the optimal solution.
Combinatorial optimization involves finding the most efficient answer among a vast number of possibilities. These problems are central to logistics, financial portfolio management, and semiconductor circuit design, but they often overwhelm traditional computing architectures as the scale of data increases.
The researchers addressed technical hurdles in previous Ising machines, such as frequency deviations between oscillators and limited connectivity. They introduced a new approach where both the oscillators and the couplers, which control interaction strength, are implemented using single silicon transistors.
By using the floating body characteristics of transistors, the team created oscillators that can have their frequencies precisely adjusted via gate voltage. This method reduces synchronization errors and allows for multi-bit coupling, which enables the system to reflect the specific weights and importance of various conditions in a complex problem.
The hardware was successfully tested on the Max-Cut problem, a representative optimization task used to maximize connections when dividing a network into two groups. Because the technology uses standard complementary metal-oxide-semiconductor (CMOS) processes, it can be mass-produced using existing South Korean semiconductor production lines without additional equipment investment.
"This research is Ising machine hardware that has secured both scalability and precision by implementing both oscillators and couplers with silicon devices," Professor Choi Yang-kyu said. "It is expected to be applied to various industrial fields requiring large-scale combinatorial optimization, such as semiconductor design automation, communication network optimization, and resource distribution."
The study, co-authored by doctoral student Yoon Seong-yun and Dr. Kim Joon-pyo, was published in the journal Science Advances on March 27, 2026.
(Reference Information)
Journal/Source: Science Advances
Title: Scalable Ising machine composed entirely of Si transistors
Link/DOI: 10.1126/sciadv.adz2384
Copyright ⓒ Aju Press All rights reserved.
