Performance Dependence of Optical-Loop-Based Coherent Ising Machine on Operating Conditions

Coherent Ising machines (CIMs) are optical computers designed to solve combinatorial optimization problems based on the Ising model. Among several variants, the optical-loop-based CIM, which utilizes optical pulses with binary phases circulating in a loop equipped with a pulse-pumped phase-sensitive amplifier (PSA) and a measurement feedback circuit, has experimentally demonstrated high performance in terms of the number of pulses and connectivity. This study numerically investigates the dependence of its calculation performance on operating conditions such as the noise level, coupling strength between pulses, and PSA pump increment rate. The CIM solving Max-Cut problems is simulated using a difference equation based on a traveling-wave model under various operating conditions. The results indicate that noise is not a critical factor, there is an optimal condition for the coupling strength, depending on the graph structure, and slower pump-increment rate leads to higher scores. Under optimal conditions, the simulation produces better calculation results than those previously reported in an experimental study.