Improved Mapping Strategies for Integrated Oscillator-Based Ising Machine Networks

Oscillator-based Ising machines (OIMs) are computational devices designed to tackle complex optimization problems by utilizing an array of interconnected oscillators. OIMs have the potential to outperform classical computing methods in both speed and energy efficiency, particularly when addressing optimization problems that prove challenging for traditional computers. One optimization problem involves mapping the nodes and edges of a random problem graph onto the Ising network. This paper aims to find a proper mapping strategy to solve this optimization problem. The proposed strategy supports large problem graph sizes and demonstrates a high success rate for medium-sized graphs, all while preserving the speed advantage inherent in oscillator-based implementations. The mapping process will be implemented using software-based methods and executed with precise visualization in MATLAB.