Gain optimization for enhanced solution quality of optoelectronic Ising machines

Area of Science:

• Physics
• Computer Science
• Engineering

Background:

• The Ising machine is a novel hardware for combinatorial optimization, offering significant speedups.
• Limited strategies exist to improve Ising machine solution quality due to reliance on physical dynamics.

Purpose of the Study:

• To systematically analyze the impact of pump gain on the optoelectronic Ising machine's solution quality.
• To develop a gain-optimization strategy for identifying optimal operating points.

Main Methods:

• Investigated the optoelectronic Ising machine's evolutionary dynamics.
• Analyzed the effects of final gain and gain control schemes.
• Proposed and applied a gain-optimization strategy.

Main Results:

• Explained experimentally observed behaviors based on system dynamics.
• Improved solution quality for Max-Cut problems across various graph densities (10%, 50%, 100%).
• Identified an optimal operating point through the proposed strategy.

Conclusions:

• The developed gain-optimization strategy enhances the performance of optoelectronic Ising machines.
• Systematic analysis of physical dynamics is key to improving Ising machine applications.
• This work advances the practical application of Ising machines for combinatorial optimization.