Pushing geophysical Ising solvers beyond straight ray tomography

Anton Ziegon, Kristopher A. Innanen

This report investigates the application of simulated Ising computations to geophysical optimization problems, building on the framework proposed by Innanen (2025). We analyze the tomographic Ising Hamiltonian for traveltime inversion via straight-ray tomography, demonstrating its ability to produce models with little to no artifacts under varying acquisition scenarios. The Ising framework is further extended to optimize seismic acquisition geometries, yielding sparser yet effective source-receiver configurations for both static straight-ray tomography as well as a timelapse FWI workflow. Additionally, we adapt the Ising machine to gravimetric inversion, introducing a method to incorporate structural constraints, and exploring joint inversion strategies to mitigate challenges commonly encountered in gravimetric inversion. While the simulated Ising machines with their optimization capabilities show promise across the considered scenarios, its sensitivity to hyperparameter tuning and current limitation to linear problems highlight the need for future research, ultimately preparing the path to non-linear Ising-based inversion, such as wavefield inversion.