The spatial photonic Ising machine (SPIM) is a simple and elegant architecture capable of expanding the spin scale with ease, and thus this optical architecture is attractive for its potential in solving large-scale combinatorial optimization problems. Here, we propose a scheme for aberration compensation to improve the computing accuracy of the SPIM based on a spatial light modulator (SLM), resulting in good scalability. First, the gamma curve of the SLM is corrected with the help of double-hole interference fringes. Then, an enhanced phase shift interferometry is used to measure the aberration of the entire optical system, and this aberration is compensated by loading the compensation phase on the SLM. Moreover, an iterative algorithm using self-adaptive cluster flipping is proposed for accelerating the convergence of SPIM. During the converging process, the number of flipped spins within the cluster is updated dynamically according to the convergence speed and accuracy, which makes it possible to achieve a good balance between high speed and high accuracy. Further, a proof-of-principle experimental set-up of the optimized SPIM is implemented, and the ground-state search of an antiferromagnetic Ising model is demonstrated. As an example, this optimized Ising machine is used for solving the number partitioning problem. It is shown that the partition of 40,000 random numbers is well demonstrated, and the final partitioning accuracy is as high as 99.88%.
KEYWORDS: Convolution, Analog electronics, Binary data, Optical computing, Computer programming, Signal processing, Optical signal processing, Computing systems, Diffractive optical elements
Here, we proposed an arbitrary hybrid analog-digital coding method which can realize high-precision digital optical convolution calculation. In this framework, the initial high-bit matrix can be decomposed into a series of low-bit encoding matrices in temporal or spatial sequence. And the final convolution result is a hybrid analog-digital matrix and obtained after decoding similar to binary-decimal conversion. As an example, we implement the analog and digital convolution calculation of two 10×10 matrices in the optical convolution calculation system. Compared with analog optical computing, this coding method is further capable of improving the optical convolutional computing accuracy of large-scale matrices.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.