Fig. 6: Estimation of time to solution distribution of chaotic amplitude control for Sherrington-Kirkpatrick (SK) instances.
From: Scaling advantage of chaotic amplitude control for high-performance combinatorial optimization
a Cumulative distribution of the time to solution P(τ) for N = 400 SK problems. The black line corresponds to the maximum of P(τ; T) with respect to the duration of the runs T. b, c, d Optimal cumulative distribution P*(τ) with \({P}^{* }(log(\tau )) \sim {{{{{{{\mathcal{N}}}}}}}}(\mu (N),\sqrt{v}(N))\) for chaotic amplitude control (CAC) (b), simulated annealing (SA) (c), and noisy mean-field annealing (NMFA) (d), respectively. e Standard deviation \(\sqrt{v}\) of the logarithm of time to solution distribution vs. problem size N. Shaded regions show the 99% confidence interval in the standard deviation. simCIM: simulation of the coherent Ising machine.
