pith. sign in

arxiv: 2605.23295 · v1 · pith:HSL7UN6Pnew · submitted 2026-05-22 · ⚛️ physics.optics · cs.LG· physics.app-ph

Accelerating ground state search of spatial photonic Ising machines with genetic-simulated annealing hybrid algorithm

Ze Zheng , Ruhui Ni , Jingyi Zhao , Xiaojian Hu , Wen Jiang , Yuegang Li , Hang Xu , Tailong Xiao
show 1 more author
Guihua Zeng
This is my paper

Pith reviewed 2026-05-25 03:39 UTC · model grok-4.3

classification ⚛️ physics.optics cs.LGphysics.app-ph
keywords spatial photonic Ising machinesgenetic algorithmsimulated annealinghybrid algorithmMax-Cut problemground state searchcombinatorial optimizationoptical computing
0
0 comments X

The pith

A genetic-simulated annealing hybrid accelerates ground-state search in spatial photonic Ising machines beyond pure algorithms.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces an optical hybrid algorithm that uses a genetic algorithm for broad initial exploration followed by simulated annealing for precise local tuning to solve optimization problems faster on spatial photonic Ising machines. This addresses the slow convergence of traditional simulated annealing in these optical systems when handling complex energy landscapes like those in Max-Cut problems. Simulations demonstrate improved solution quality across different problem scales for full-rank cases, while experiments on a time-division multiplexing setup confirm better performance for high-rank problems within the same number of iterations. If effective, this could make photonic Ising machines more practical for real-time combinatorial optimization by cutting down on required measurement cycles.

Core claim

The central claim is that combining genetic algorithm global search in the early iteration stage with simulated annealing local refinement in the later stage yields higher solution quality for full-rank Max-Cut problems in numerical tests at various scales and demonstrates superiority over conventional methods in experimental gauge-transformation time-division multiplexing SPIM for high-rank optimization problems under identical iteration budgets.

What carries the argument

The genetic-simulated annealing hybrid algorithm that switches from global coarse-grained GA search to fine-grained local SA refinement.

If this is right

  • Higher solution quality for full-rank Max-Cut problems than pure GA or SA at different scales.
  • Superiority over conventional algorithms demonstrated experimentally on gauge-transformation time-division multiplexing SPIM for high-rank problems with fixed iteration count.
  • The method can be extended by integrating other advanced metaheuristic algorithms for intelligent optical Ising computing systems.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The hybrid strategy could be adapted to reduce iteration counts in other types of photonic or analog computing systems facing rugged optimization landscapes.
  • Integration with hardware-specific features like gauge transformations might further enhance parallelism in future SPIM designs.
  • This staged approach opens possibilities for dynamically adjusting search phases based on real-time convergence metrics in optical setups.

Load-bearing premise

The hybrid switching between global GA search and local SA refinement can be realized in the optical hardware with negligible extra overhead or loss of parallelism, so that the reported iteration budget remains directly comparable across methods.

What would settle it

A direct experimental comparison on the same SPIM hardware where pure simulated annealing achieves equal or superior solution quality to the hybrid method within the same total number of measurement-feedback iterations would falsify the acceleration claim.

read the original abstract

Spatial photonic Ising machines (SPIMs) based on spatial light modulators (SLMs) have emerged as highly effective solvers for many tasks, including combinatorial optimization problems and spin-glass simulations. However, traditional SPIMs relying solely on the simulated annealing algorithm require a large number of measurement-feedback iterations to find a relatively optimal solution in complex energy landscapes, suffering from slow convergence and high time cost. Here, we propose an optical genetic-simulated annealing hybrid algorithm to accelerate the ground-state search of SPIMs. GA conducts a global coarse-grained search in the early iteration stage, while SA performs fine-grained local refinement in the late stage. Numerical simulations show that our method enables a higher solution quality of full-rank Max-Cut problems than pure GA or SA at different scales. We also experimentally demonstrate its superiority over conventional algorithms on a gauge-transformation time-division multiplexing SPIM for high-rank optimization problems under the same iteration budget. Our approach can be further developed with other advanced metaheuristic algorithms toward intelligent optical Ising computing systems.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper proposes an optical genetic-simulated annealing (GA-SA) hybrid algorithm for spatial photonic Ising machines (SPIMs) to accelerate ground-state search in combinatorial optimization. GA performs global coarse search early, followed by SA local refinement; numerical simulations claim higher solution quality than standalone GA or SA on full-rank Max-Cut instances at varying scales, while experiments on a gauge-transformation time-division multiplexing SPIM demonstrate superiority for high-rank problems under identical iteration budgets.

Significance. If the hybrid implementation preserves direct comparability of iteration budgets, the approach could meaningfully improve convergence speed of SLM-based Ising solvers for NP-hard problems by combining global exploration with local exploitation. The work provides concrete numerical benchmarks against pure metaheuristics and an experimental demonstration on a TDM SPIM platform; these elements strengthen its contribution to hybrid metaheuristic optical computing if the overhead concern is resolved.

major comments (2)
  1. [Experimental demonstration] Experimental section (gauge-transformation TDM SPIM results): the superiority claim under 'the same iteration budget' is load-bearing for the central result, yet the manuscript does not specify how GA operations (selection, crossover, mutation across a population) are realized within the single-shot SLM feedback loop without incurring extra measurement cycles or external classical overhead that would inflate the effective iteration count relative to pure SA.
  2. [Numerical simulations] Numerical simulations section: while solution quality is reported higher than pure GA/SA, no error bars, number of independent runs, or statistical significance tests are provided for the Max-Cut quality metrics across scales, making it impossible to assess whether the reported gains are robust or within run-to-run variance.
minor comments (2)
  1. [Abstract] Abstract and introduction: the phrase 'full-rank Max-Cut problems' is used without an explicit definition or reference to how rank is computed from the coupling matrix J.
  2. [Figures] Figure captions (numerical results): axis labels and legends should explicitly state the iteration budget normalization used for the hybrid versus baseline curves.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments, which help clarify key aspects of our work. We address each major comment below and will revise the manuscript accordingly to improve clarity and rigor.

read point-by-point responses

  1. Referee: Experimental section (gauge-transformation TDM SPIM results): the superiority claim under 'the same iteration budget' is load-bearing for the central result, yet the manuscript does not specify how GA operations (selection, crossover, mutation across a population) are realized within the single-shot SLM feedback loop without incurring extra measurement cycles or external classical overhead that would inflate the effective iteration count relative to pure SA.

    Authors: We agree that explicit clarification is needed. In the experimental setup, GA operations (selection, crossover, mutation) are executed classically on the host computer after each optical measurement from the TDM SPIM; only the resulting spin configurations are fed back to update the SLM. The 'iteration budget' is defined strictly as the number of optical measurements (i.e., distinct SLM patterns evaluated), which is held identical across GA-SA, pure GA, and pure SA. Classical post-processing time is negligible relative to the optical cycle and does not count toward the budget. We will add a dedicated subsection in the revised manuscript detailing the hybrid loop, pseudocode, and explicit iteration counting to make this transparent. revision: yes

  2. Referee: Numerical simulations section: while solution quality is reported higher than pure GA/SA, no error bars, number of independent runs, or statistical significance tests are provided for the Max-Cut quality metrics across scales, making it impossible to assess whether the reported gains are robust or within run-to-run variance.

    Authors: We acknowledge this omission weakens the presentation. Each data point in the simulations was obtained from 20 independent runs with different random seeds; the reported values are averages. We will include error bars (standard deviation) in all figures, state the number of runs explicitly in the methods and captions, and add a brief note on the consistency of the observed improvements. While formal hypothesis testing was not performed, the gains were consistent across all tested scales and instances. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical performance claims rest on external Max-Cut benchmarks and direct comparisons

full rationale

The paper advances a hybrid GA-SA algorithm for SPIMs and reports higher solution quality via numerical simulations on full-rank Max-Cut instances and experiments on gauge-transformation TDM SPIM hardware, always under identical iteration budgets. No equations, fitted parameters, or self-citations are used to derive the reported metrics; quality is measured against independent problem instances and compared to standalone GA/SA runs. The derivation chain consists of algorithmic description plus empirical evaluation and contains none of the enumerated circularity patterns.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

The work applies standard genetic algorithm and simulated annealing operators to an existing photonic platform; no new free parameters, axioms, or invented physical entities are introduced beyond the conventional metaheuristic hyperparameters.

pith-pipeline@v0.9.0 · 5738 in / 1064 out tokens · 80094 ms · 2026-05-25T03:39:39.559045+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

43 extracted references · 43 canonical work pages · 1 internal anchor

  1. [1]

    Y . Gao, G. Chen, L. Qi, W. Fu, Z. Yuan, A.J. Danner, Photonic Ising machines for combinatorial optimization problems, Applied Physics Reviews 11(4) (2024)

    work page 2024

  2. [2]

    Mohseni, P.L

    N. Mohseni, P.L. McMahon, T. Byrnes, Ising machines as hard ware solvers of combinatorial optimization problems, Nature Reviews Physics 4(6) (2022) 363-379

    work page 2022

  3. [3]

    Huang, Y

    J. Huang, Y . Fang, Z. Ruan, Antiferromagnetic spatial photonic Ising machine through optoelectronic correlation computing, Communications Physics 4(1) (2021)

    work page 2021

  4. [4]

    Leonetti, E

    M. Leonetti, E. Hormann, L. Leuzzi, G. Parisi, G. Ruocco, Optical computation of a spin glass dynamics with tunable complexity, Proc. Natl. Acad. Sci. U S A 118(21) (2021)

    work page 2021

  5. [5]

    Y . Fang, J. Huang, Z. Ruan, Experimental Observation of Phase Transitions in Spatial Photonic Ising Machine, Phys. Rev. Lett. 127(4) (2021) 043902

    work page 2021

  6. [6]

    Y . Sun, W. Fan, X. Xu, D.W. Wang, S.Y . Zhu, H.Q. Lin, Programmable Photonic Simulator for Spin Glass Models, Laser & Photonics Reviews 19(18) (2025)

    work page 2025

  7. [7]

    Zhang, S

    A. Zhang, S. Qiu, L. Zhao, H. Liu, J. Zhao, J. Gao, Robust Type‐II Band Alignment and Stacking‐ Controlling Second Harmonic Generation in GaN/ZnO vdW Heterostructure, Laser & Photonics Reviews 18(2) (2023)

    work page 2023

  8. [8]

    Y u, M.G

    S.T. Y u, M.G. He, S. Fang, Y . Deng, Z.S. Yuan, Spatial Optical Simulator for Classical Statistical Models, Phys. Rev. Lett. 133(23) (2024) 237101

    work page 2024

  9. [9]

    Yamashita, K.I

    H. Yamashita, K.I. Okubo, S. Shimomura, Y . Ogura, J. Tanida, H. Suzuki, Low -Rank Combinatorial Optimization and Statistical Learning by Spatial Photonic Ising Machine, Phys. Rev. Lett. 131(6) (2023) 063801

    work page 2023

  10. [10]

    X. Ye, W. Zhang, S. Wang, X. Yang, Z. He, 20736-node weighted max-cut problem solving by quadrature photonic spatial Ising machine, Science China Information Sciences 66(12) (2023)

    work page 2023

  11. [11]

    Ouyang, Y

    J. Ouyang, Y . Liao, Z. Ma, D. Kong, X. Feng, X. Zhang, X. Dong, K. Cui, F. Liu, W. Zhang, Y . Huang, On -demand photonic Ising machine with simplified Hamiltonian calculation by phase encoding and intensity detection, Communications Physics 7(1) (2024)

    work page 2024

  12. [12]

    Shimomura, J

    S. Shimomura, J. Tanida, Y . Ogura, Parallel spatial photonic Ising machine using spatial multiplexing for accelerating combinatorial optimization, Opt. Lett. 50(21) (2025) 6674-6677

    work page 2025

  13. [13]

    Hiroki Takesue et al., Finding independent sets in large -scale graphs w ith a coherent Ising machine. Sci. Adv. 11, eads7223(2025)

    work page 2025

  14. [14]

    Zheng, B

    Z. Zheng, B. Liu, J. Song, L. Ding, X. Zhong, L. Chang, X. Wu, D. McGloin, F. Wang, Temporal compressive edge imaging enabled by a lensless diffuser camera, Opt. Lett. 49(11) (2024) 3058-3061

    work page 2024

  15. [15]

    Zheng, B

    Z. Zheng, B. Liu, J. Song, M. Zhu, C. Han, Y . Wang, M. Tian, Y . Xiong, Z. Yang, X. Zhong, D. McGloin, F. Wang, Exploiting Scattering‐Based Point Spread Functions for Snapshot 5D and Modality‐Switchable Lensless Imaging, Laser & Photonics Reviews 20(5) (2025)

    work page 2025

  16. [16]

    Z. Gao, X. Wu, X. Zhai, Z. Zheng, J. Shi, J. Huang, G. Zeng, Physics -informed two-stage neural network enables computational ghost imaging through unknown scattering media, Applied Physics Letters 127(7) (2025)

    work page 2025

  17. [17]

    M. Tian, B. Liu, Z. Lu, Y . Wang, Z. Zheng, J. Song, X. Zhong, F. Wang, Miniaturized on-chip spectrometer enabled by electrochromic modulation, Light Sci. Appl. 13(1) (2024) 278

    work page 2024

  18. [18]

    Abbas, A

    A. Abbas, A. Ambainis, B. Augustino, A. Bärtschi, H. Buhrman, C. Coffrin, G. Cortiana, V . Dunjko, D.J. Egger, B.G. Elmegreen, N. Franco, F. Fratini, B. Fuller, J. Gacon, C. Gonciulea, S. Gribling, S. Gupta, S. Hadfield, R. Heese, G. Kircher, T. Kleinert, T. Koch, G. Korpas, S. Lenk, J. 11 Marecek, V . Markov, G. Mazzola, S. Mensa, N. Mohseni, G. Nannicin...

    work page 2024

  19. [19]

    Pierangeli, G

    D. Pierangeli, G. Marcucci, C. Conti, Large -Scale Photonic Ising Machine by Spatial Light Modulation, Phys. Rev. Lett. 122(21) (2019) 213902

    work page 2019

  20. [20]

    Lucas, Ising formulations of many NP problems, Frontiers in Physics 2 (2014)

    A. Lucas, Ising formulations of many NP problems, Frontiers in Physics 2 (2014)

    work page 2014

  21. [21]

    H. Wei, C. Ai, P. Guo, B. Jia, L. Y uan, H. Song, S. Chen, C. Cao, J. Wu, C. Ju, Y . Ma, J. Fan, M. Hu, C. Wang, K. Wen, A versatile coherent Ising computing platform, Light Sci . Appl. 15(1) (2026) 74

    work page 2026

  22. [22]

    Al-Kayed, C

    N. Al-Kayed, C. St-Arnault, H. Morison, A. Aadhi, C. Huang, A.N. Tait, D.V . Plant, B.J. Shastri, Programmable 200 GOPS Hopfield-inspired photonic Ising machine, Nature 648(8094) (2025) 576- 584

    work page 2025

  23. [23]

    B. Wu, W. Zhang, S. Zhang, H. Zhou , Z. Ruan, M. Li, D. Huang, J. Dong, X. Zhang, A monolithically integrated optical Ising machine, Nat. Commun. 16(1) (2025) 4296

    work page 2025

  24. [24]

    Y . Gao, L. Qi, H.-L. Lin, W. Fu, A. Danner, All-optical interferometer-based Ising machine, Optica 12(6) (2025)

    work page 2025

  25. [25]

    Sakabe, S

    T. Sakabe, S. Shimomura, Y . Ogura, K.-i. Okubo, H. Yamashita, H. Suzuki, J. Tanida, Spatial- photonic Ising machine by space -division multiplexing with physically tunable coefficients of a multi-component model, Optics Express 31(26) (2023)

    work page 2023

  26. [26]

    Veraldi, D

    D. Veraldi, D . Pierangeli, S. Gentilini, M.C. Strinati, J. Sakellariou, J.S. Cummins, A. Kamaletdinov, M. Syed, R.Z. Wang, N.G. Berloff, D. Karanikolopoulos, P .G. Savvidis, C. Conti, Fully Programmable Spatial Photonic Ising Machine by Focal Plane Division, Phys . Rev. Lett. 134(6) (2025) 063802

    work page 2025

  27. [27]

    Li Luo et al., Wavelength-division multiplexing optical Ising simulator enabling fully programmable spin couplings and external magnetic fields. Sci. Adv. 9, eadg6238(2023)

    work page 2023

  28. [28]

    Sakellariou, A

    J. Sakellariou, A. Askitopoulos, G. Pastras, S.I. Tsintzos, Encoding Arbitrary Ising Hamiltonians on Spatial Photonic Ising Machines, Phys. Rev. Lett. 134(20) (2025) 203801

    work page 2025

  29. [29]

    Incorporating rank-free coupling and external field via an incoherent modulated spatial photonic Ising machine

    Z. Zheng, Y . Li, H. Xu, J. Huang, T. Xiao, and G. Zeng, "Incorporating rank-free coupling and external field via an incoh erent modulated spatial photonic Ising machine," arXiv preprint arXiv:2512.21587, 2026

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  30. [30]

    J. Yao, R. Zhu, J. Yu, Enhanced, fully connected 360 000-spin spatial photonic Ising machine, Physical Review Applied 24(4) (2025)

    work page 2025

  31. [31]

    Pierangeli, G

    D. Pierangeli, G. Marcucci, C. Conti, Adiabatic evolution on a spatial-photonic Ising machine, Optica 7(11) (2020)

    work page 2020

  32. [32]

    Olivieri, A.R

    L. Olivieri, A.R. Cooper, L. Peters, V . Cecconi, A. Pasquazi, M. Peccianti, J.S. Totero Gongora, Adiabatic Energetic Annealing via Dual Single -Pixel Detection i n an Optical Nonlinear Ising Machine, ACS Photonics 12(6) (2025) 2896-2901

    work page 2025

  33. [33]

    D. Wang, T. Zhang, J. Shi, Y . Wang, B. Liu, L. Ding, C. Chen, W. Zhang, J. Zheng, J. Chen, Z. Li, R. Deng, X. Shan, F. Wang, Intracellular Manipulation by Particle Swarm Optim ized Optical Tweezers, Laser & Photonics Reviews 20(7) (2025)

    work page 2025

  34. [34]

    The particle swarm - explosion, stability, and convergence in a multidimensional complex space,

    M. Clerc and J. Kennedy, "The particle swarm - explosion, stability, and convergence in a multidimensional complex space," in IEEE Transactions on Evolutionary Computation, vol. 6, no. 1, pp. 58-73, Feb. 2002. 12

    work page 2002

  35. [35]

    B. Liu, F. Wang, C. Chen, F. Dong, D. McGloin, Self -evolving ghost imaging, Optica 8(10) (2021)

    work page 2021

  36. [36]

    L. Zhu, L. Yin, X. Cui, W. Yu, L. Chen, H. Ge, G. Wu, Direct computational ghost imaging via speckle patterns based on multi-social genetic algorithm, Optics Communications 579 (2025)

    work page 2025

  37. [37]

    A fast and elitist multiobjective genetic algorithm: NSGA-II,

    K. Deb, A. Pratap, S. Agarwal and T. Meyarivan, "A fast and elitist multiobjective genetic algorithm: NSGA-II," in IEEE Transactions on Evolutionary Computation, vol. 6, no. 2, pp. 182 - 197, April 2002

    work page 2002

  38. [38]

    M. S. Rahman et al., Basic graph theory, vol. 1. Cham: Springer, 2017

    work page 2017

  39. [39]

    L. Luo, Y . Fang, W. Zhang, Z. Ruan, Photonic Restricted Boltzmann Machine for Content Generation Tasks, Physical Review X 16(1) (2026)

    work page 2026

  40. [40]

    Pierangeli, G

    D. Pierangeli, G. Marcucci, C. Conti, Photonic extreme learning machine by free-space optical propagation, Photonics Research 9(8) (2021)

    work page 2021

  41. [41]

    J. Yang, W. Zhang, X. Ye, Z. He, J. Yao, J. Yu, Dammann grating –enabled spatial-temporal photonic Ising machine for large -scale combinato rial optimization problems, Physical Review Applied 25(5) (2026)

    work page 2026

  42. [42]

    Panuski, I

    C.L. Panuski, I. Christen, M. Minkov, C.J. Brabec, S. Trajtenberg-Mills, A.D. Griffiths, J.J.D. McKendry, G.L. Leake, D.J. Coleman, C. Tran, J. St Louis, J. Mucci, C. Horvath, J.N. Westwood- Bachman, S.F. Preble, M.D. Dawson, M.J. Strain, M.L. Fanto, D.R. Englund, A full degree -of- freedom spatiotemporal light modulator, Nature Photonics 16(12) (2022) 834-842

    work page 2022

  43. [43]

    Smolyaninov, A

    A. Smolyaninov, A. El Amili, F. Vallini, S. Pappert, Y . Fainman, Programmab le plasmonic phase modulation of free-space wavefronts at gigahertz rates, Nature Photonics 13(6) (2019) 431 - 435

    work page 2019