Analyzing Parametric Oscillator Ising Machines through the Kuramoto Lens

TL;DR

The work addresses the gap between parametric-oscillator Ising machines (PoIMs) and traditional oscillator-based Ising machines (OIMs) by deriving a canonical phase description from a conjugate Stuart-Landau model pumped at $2\omega$. It reveals a Kuramoto-like phase dynamics with both phase-difference and phase-sum couplings, supported by an energy function that maps to the Ising Hamiltonian in the binary-phase limit. A key finding is that explicit second-harmonic driving is unnecessary in PoIMs, and that amplitude heterogeneity rescales spin interactions, potentially degrading solution quality. Together, these results provide a unified analytical framework for designing and understanding oscillator-based Ising solvers, enabling cross-pollination of strategies across PoIM and OIM architectures.

Abstract

Networks of coupled nonlinear oscillators are emerging as powerful physical platforms for implementing Ising machines. Yet the relationship between parametric-oscillator implementations and traditional oscillator-based Ising machines remains underexplored. In this work, we develop a Kuramoto-style, canonical phase description of parametric oscillator Ising machines by starting from the Stuart-Landau oscillator model -- the canonical normal form near a Hopf bifurcation, and a natural reduced description for many parametric oscillator implementations such as the degenerate optical parametric oscillator (DOPO) among others. The resulting phase dynamics combine the usual phase-difference coupling observed in the standard Kuramoto model along with an intrinsic phase sum term that is generated when conjugate coupling is considered. Moreover, our formulation helps explain why explicit second-harmonic driving is unnecessary in parametric oscillators and also reveals how quasi-steady amplitude heterogeneity scales the original strength of the spin interaction with potentially adverse impacts on the solution quality. Our work helps develop a unifying view of the oscillator-based approach to designing Ising machines.

Figures (3)

• Figure 1: MaxCut solutions obtained using Kuramoto-based OIM dynamics and the canonical phase form of the PoIM dynamics for random generated graphs of varying sizes: (a) $N = 50$; (b) $N = 100$; (c) $N = 150$. For each graph size, 25 independent graph instances are considered, and each instance is simulated 10 times using both models. The best solution from each model is used for comparison.
• Figure 2: Evolution of the lowest Ising energy obtained across varying levels of amplitude heterogeneity (AH) for an illustrative 50-node regular graph with degree 5. The results corroborate the theoretical result that increasing AH degrades solution quality. Trials=100.
• Figure 3: (a) Illustrative graph with 15 nodes and 58 edges. Evolution of the observable output in (b) Stationary; and (c) Rotating frames of reference.