pith. sign in

arxiv: 2209.10814 · v3 · submitted 2022-09-22 · 🧮 math.NA · cs.NA

An Alternating Direction Method of Multipliers for Inverse Lithography Problem

Junqing Chen , Haibo Liu This is my paper

Pith reviewed 2026-05-24 11:06 UTC · model grok-4.3

classification 🧮 math.NA cs.NA
keywords inverse lithographyADMMmask optimizationtotal variation regularizationthreshold truncationaugmented Lagrangian
0
0 comments X

The pith

An ADMM algorithm solves inverse lithography optimization by splitting it into subproblems with an analytical solution for the imaging term.

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

The paper proposes using the alternating direction method of multipliers to minimize an objective that measures how well a mask produces a target wafer pattern, while penalizing non-binary masks and adding total variation regularization. Variable splitting creates an augmented Lagrangian whose subproblems each admit efficient solutions. The imaging subproblem is handled directly by threshold truncation instead of iterative approximation, and the overall iteration is shown to converge.

Core claim

By variable splitting the misfit, binary penalty, and total variation terms, the augmented Lagrangian framework yields subproblems that can be solved efficiently; the imaging subproblem is solved analytically via threshold truncation of the imaging function, and the resulting ADMM iteration converges for the inverse lithography problem.

What carries the argument

Alternating direction method of multipliers applied after variable splitting on the three-term objective, with the imaging subproblem solved by direct threshold truncation.

If this is right

  • The binary mask constraint and total variation term can be enforced without solving a full non-smooth optimization at each step.
  • The imaging step requires only a simple truncation operation rather than repeated sigmoid evaluations.
  • Convergence guarantees apply directly to the split formulation used for lithography mask design.
  • Numerical examples confirm that the decomposed iteration produces usable masks for standard test patterns.

Where Pith is reading between the lines

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

  • The same splitting pattern could be tested on other inverse imaging problems that combine fidelity, binarity, and edge-preserving penalties.
  • Runtime comparisons against gradient-based or level-set methods on the same mask targets would quantify the practical speedup.
  • If the threshold truncation step generalizes, the method may extend to related binary design tasks in optics or materials.

Load-bearing premise

The variable splitting and augmented Lagrangian produce subproblems whose solutions remain efficient and cover the original objective under the specific constraints and regularizers used.

What would settle it

A concrete lithography mask design instance in which the ADMM iterates fail to reduce the misfit below a known lower bound or produce a non-convergent sequence.

Figures

Figures reproduced from arXiv: 2209.10814 by Haibo Liu, Junqing Chen.

Figure 1
Figure 1. Figure 1: Lithography system resist on wafer, this optical lithography process can transfer the mask pattern to wafer pattern for next stage use. Under the assumption that the mask is thin, the projection processing can be described by Fourier Optics [2]. Using the Kirchhoff approximation, with the help of Abbe formulation, the aerial image on the wafer with partially coherent illumination can be approximately descr… view at source ↗
Figure 2
Figure 2. Figure 2: Point spread function for lens with circular aperture. [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: The curves of Sigmoid function and its derivative with [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: The i-th entry of ∇ha(V ) and i-th diagonal entry of Hessian of ha(V ) with a = 20, tr = 0.3. The left is with Ii = 0 and the right is with Ii = 1 . We begin our analysis by showing a sufficient decrease property of the La￾grange function. Lemma 3.2. If ρ 2 − lh ρ − lh > 0, where lh is the Lipschitz constant of ∇h, there is a constant C = C(ρ, lh) > 0 such that for sufficiently large k, we have L(U k , V k… view at source ↗
Figure 5
Figure 5. Figure 5: The curves of objective function with Wk i = 0.2, ρ = 1, tr = 0.3. The left is with Ii = 1, the right is with Ii = 0. 4.1 Choosing parameters In this example, we will discuss the influence of the parameters ρ, β1, β2 in (16) and γ in (28). Figure (6) shows the target image in this example. The size of the image is 144×144 and the blue part is valued 0, and the yellow represents value 1. Figure (7) shows er… view at source ↗
Figure 6
Figure 6. Figure 6: The target pattern squares and long strips. For this kind of mixed pattern, the algorithm still works well. From the second and forth rows, the small details can be imaged with optimized mask patterns, even for imaging system with defocus. 5 Conclusion We have developed an ADMM method to solve the inverse lithography problem. The TV regularization is introduced to deal with ill-posedness of the inverse pro… view at source ↗
Figure 7
Figure 7. Figure 7: Error decreasing with respect to ρ, with γ = 50, β1 = 0.005, β2 = 0.01 [4] Y. Peng, J. Zhang, Y. Wang, and Z. Yu, “Gradient-based source and mask optimization in optical lithography,” IEEE Transactions on image process￾ing, vol. 20, no. 10, pp. 2856–2864, 2011. 1 [5] X. Ma and G. R. Arce, “Pixel-based OPC optimization based on conjugate gradients,” Optics express, vol. 19, no. 3, pp. 2165–2180, 2011. 1 [6]… view at source ↗
Figure 8
Figure 8. Figure 8: A heuristic choice of γ, with fixed ρ = 10, β1 = 0.005, β2 = 0.01. (a) 20 40 60 80 100 120 140 20 40 60 80 100 120 140 (b) 20 40 60 80 100 120 140 20 40 60 80 100 120 140 (c) 20 40 60 80 100 120 140 20 40 60 80 100 120 140 [PITH_FULL_IMAGE:figures/full_fig_p024_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: The optimal mask with different β1. From left to right β1 = 0.005, 0.01, 0.015. [12] Y. Shen, F. Peng, and Z. Zhang, “Semi-implicit level set formulation for lithographic source and mask optimization,” Optics Express, vol. 27, no. 21, pp. 29 659–29 668, 2019. 1 [13] A. Erdmann, T. Fuehner, T. Schnattinger, and B. Tollkuehn, “Toward automatic mask and source optimization for optical lithography,” in Optical… view at source ↗
Figure 10
Figure 10. Figure 10: The output pattern with different β2. IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, vol. 39, no. 10, pp. 2822–2834, 2019. 1 [16] H. Yang, Z. Li, K. Sastry, S. Mukhopadhyay, M. Kilgard, A. Anand￾kumar, B. Khailany, V. Singh, and H. Ren, “Generic lithography mod￾eling with dual-band optics-inspired neural networks,” arXiv preprint arXiv:2203.08616, 2022. 1 [17] L. Pang, “Inv… view at source ↗
Figure 11
Figure 11. Figure 11: Left column: mask, Middle column: output pattern, Right column: [PITH_FULL_IMAGE:figures/full_fig_p026_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Left column: mask, Middle column: output pattern, Right column: [PITH_FULL_IMAGE:figures/full_fig_p027_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Left column: mask, Middle column: output pattern, Right column: [PITH_FULL_IMAGE:figures/full_fig_p028_13.png] view at source ↗
read the original abstract

We propose an alternating direction method of multipliers (ADMM) to solve an optimization problem stemming from inverse lithography. The objective functional of the optimization problem includes three terms: the misfit between the imaging on wafer and the target pattern, the penalty term which ensures the mask is binary and the total variation regularization term. By variable splitting, we introduce an augmented Lagrangian for the original objective functional. In the framework of ADMM method, the optimization problem is divided into several subproblems. Each of the subproblems can be solved efficiently. We give the convergence analysis of the proposed method. Specially, instead of solving the subproblem concerning sigmoid, we solve directly the threshold truncation imaging function which can be solved analytically. We also provide many numerical examples to illustrate the effectiveness of the method.

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 / 1 minor

Summary. The manuscript proposes an ADMM algorithm for an inverse lithography optimization problem whose objective combines an imaging misfit term, a binary-mask penalty, and total-variation regularization. Variable splitting produces an augmented Lagrangian that is minimized by alternating subproblem solves; the imaging subproblem is replaced by direct threshold truncation (claimed to be analytically solvable) rather than a sigmoid, and a convergence analysis together with numerical examples is supplied.

Significance. If the convergence result can be shown to apply to the threshold-truncation variant, the method would supply a practical, splitting-based solver for a class of non-smooth ILT problems that is currently handled by more expensive gradient or heuristic approaches.

major comments (2)
  1. [Abstract / convergence section] Abstract (and the convergence-analysis section referenced therein): the statement that 'the convergence analysis of the proposed method' is given is load-bearing, yet the implemented algorithm replaces the sigmoid imaging subproblem by a hard threshold truncation operator. Standard ADMM convergence arguments rely on convexity or Lipschitz continuity of the original proximal mappings; the paper must explicitly verify that the analysis carries over to the non-differentiable, set-valued threshold operator under the binary penalty and TV terms, or supply a separate proof.
  2. [Method description (variable splitting and subproblem derivation)] The claim that 'each of the subproblems can be solved efficiently' and that the imaging subproblem 'can be solved analytically' via threshold truncation is central to the contribution, but no derivation is supplied showing how the augmented-Lagrangian subproblem for the imaging term reduces exactly to a simple threshold operation once the binary and TV terms are split.
minor comments (1)
  1. Notation for the imaging operator, the sigmoid approximation, and the threshold truncation should be introduced with explicit functional definitions before they are used in the algorithm statement.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and constructive comments on our manuscript. We address the major comments point by point below.

read point-by-point responses

  1. Referee: [Abstract / convergence section] Abstract (and the convergence-analysis section referenced therein): the statement that 'the convergence analysis of the proposed method' is given is load-bearing, yet the implemented algorithm replaces the sigmoid imaging subproblem by a hard threshold truncation operator. Standard ADMM convergence arguments rely on convexity or Lipschitz continuity of the original proximal mappings; the paper must explicitly verify that the analysis carries over to the non-differentiable, set-valued threshold operator under the binary penalty and TV terms, or supply a separate proof.

    Authors: We thank the referee for this observation. The convergence analysis in the manuscript applies to the ADMM iterates with the specific proximal mappings used, including the threshold truncation (which is the proximal operator of the indicator function of the binary set and hence firmly nonexpansive). The other terms (binary penalty and TV) are also handled via proximal operators of convex functions, satisfying the standard conditions for ADMM convergence. To make the applicability explicit for the threshold operator, we will add a clarifying paragraph in the convergence section of the revised manuscript. revision: partial

  2. Referee: [Method description (variable splitting and subproblem derivation)] The claim that 'each of the subproblems can be solved efficiently' and that the imaging subproblem 'can be solved analytically' via threshold truncation is central to the contribution, but no derivation is supplied showing how the augmented-Lagrangian subproblem for the imaging term reduces exactly to a simple threshold operation once the binary and TV terms are split.

    Authors: We agree that an explicit derivation strengthens the paper. In the revised manuscript we will insert a step-by-step derivation of the imaging subproblem: after splitting, the relevant augmented-Lagrangian term decouples into a quadratic misfit plus a quadratic penalty on the auxiliary variable; the minimizer is obtained exactly by applying the threshold truncation operator (setting the imaging variable to 1 or 0 according to whether the adjusted target exceeds the threshold determined by the penalty parameter and multiplier). This derivation will be placed immediately after the variable-splitting step. revision: yes

Circularity Check

0 steps flagged

No circularity; derivation applies standard ADMM to stated objective with explicit modifications

full rationale

The paper states an objective with misfit, binary penalty, and TV terms; introduces variable splitting and augmented Lagrangian; divides into subproblems solved efficiently (imaging via direct threshold truncation instead of sigmoid); and separately states convergence analysis for the resulting algorithm. No equation reduces to an input by construction, no parameter is fitted then renamed as prediction, and no load-bearing claim rests on self-citation or imported uniqueness. The threshold truncation is presented as an explicit replacement for analytical solvability, with the analysis asserted to cover the implemented method. This is self-contained against external benchmarks (standard ADMM theory plus numerical validation).

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The paper relies on standard mathematical properties of ADMM convergence and the standard formulation of the inverse lithography objective from the lithography literature; no new entities or fitted parameters are introduced in the abstract.

axioms (1)
  • standard math Convergence of ADMM iterations under suitable conditions on the objective and splitting
    Invoked when stating the convergence analysis of the proposed method.

pith-pipeline@v0.9.0 · 5656 in / 1042 out tokens · 59287 ms · 2026-05-24T11:06:40.922660+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. [1]

    Inverse lithography technology (ILT): What is the impact to the photomask industry?

    L. Pang, Y. Liu, and D. Abrams, “Inverse lithography technology (ILT): What is the impact to the photomask industry?” in Photomask and Next- Generation Lithography Mask Technology XIII , vol. 6283. International Society for Optics and Photonics, 2006, p. 62830X. 1

    work page 2006

  2. [2]

    Introduction to fourier optics,

    J. W. Goodman, “Introduction to fourier optics,” Introduction to Fourier optics, 3rd ed., by JW Goodman. Englewood, CO: Roberts & Co. Publishers, 2005, vol. 1, 2005. 1, 2.1, 2.1

    work page 2005

  3. [3]

    Fast pixel-based mask optimization for inverse lithography,

    Y. Granik, “Fast pixel-based mask optimization for inverse lithography,” Journal of Micro/Nanolithography, MEMS, and MOEMS , vol. 5, no. 4, p. 043002, 2006. 1 22 0 20 40 60 80 100 120 140 iteration step 0 500 1000 1500 2000 2500 3000error =5 =10 =15 =20 Figure 7: Error decreasing with respect to ρ, with γ = 50,β 1 = 0.005,β 2 = 0.01

    work page 2006

  4. [4]

    Gradient-based source and mask optimization in optical lithography,

    Y. Peng, J. Zhang, Y. Wang, and Z. Yu, “Gradient-based source and mask optimization in optical lithography,” IEEE Transactions on image process- ing, vol. 20, no. 10, pp. 2856–2864, 2011. 1

    work page 2011

  5. [5]

    Pixel-based OPC optimization based on conjugate gradients,

    X. Ma and G. R. Arce, “Pixel-based OPC optimization based on conjugate gradients,” Optics express, vol. 19, no. 3, pp. 2165–2180, 2011. 1

    work page 2011

  6. [6]

    Inverse image problem of designing phase shifting masks in optical lithography,

    S. H. Chan and E. Y. Lam, “Inverse image problem of designing phase shifting masks in optical lithography,” in 2008 15th IEEE International Conference on Image Processing. IEEE, 2008, pp. 1832–1835. 1

    work page 2008

  7. [7]

    Initialization for robust inverse synthesis of phase-shifting masks in optical projection lithography,

    S. H. Chan, A. K. Wong, and E. Y. Lam, “Initialization for robust inverse synthesis of phase-shifting masks in optical projection lithography,” Optics Express, vol. 16, no. 19, pp. 14 746–14 760, 2008. 1

    work page 2008

  8. [8]

    Generalized inverse lithography methods for phase- shifting mask design,

    X. Ma and G. R. Arce, “Generalized inverse lithography methods for phase- shifting mask design,” Optics Express, vol. 15, no. 23, pp. 15 066–15 079,

    work page

  9. [9]

    Double-exposure mask synthesis us- ing inverse lithography,

    A. A. Poonawala and P. Milanfar, “Double-exposure mask synthesis us- ing inverse lithography,” Journal of Micro/Nanolithography, MEMS, and MOEMS, vol. 6, no. 4, p. 043001, 2007. 1

    work page 2007

  10. [10]

    Model-based assist feature placement: an inverse imaging approach,

    A. Poonawala, B. Painter, and J. Mayhew, “Model-based assist feature placement: an inverse imaging approach,” in Photomask Technology 2008, vol. 7122. SPIE, 2008, pp. 277–286. 1

    work page 2008

  11. [11]

    Level-set-based inverse lithography for photomask synthesis,

    Y. Shen, N. Wong, and E. Y. Lam, “Level-set-based inverse lithography for photomask synthesis,” Optics Express, vol. 17, no. 26, pp. 23 690–23 701,

    work page

  12. [12]

    1 23 Figure 8: A heuristic choice of γ, with fixed ρ = 10,β 1 = 0.005,β 2 = 0.01. (a) 20 40 60 80 100 120 140 20 40 60 80 100 120 140 (b) 20 40 60 80 100 120 140 20 40 60 80 100 120 140 (c) 20 40 60 80 100 120 140 20 40 60 80 100 120 140 Figure 9: The optimal mask with different β1. From left to right β1 = 0.005, 0.01, 0.015

    work page

  13. [13]

    Semi-implicit level set formulation for lithographic source and mask optimization,

    Y. Shen, F. Peng, and Z. Zhang, “Semi-implicit level set formulation for lithographic source and mask optimization,” Optics Express, vol. 27, no. 21, pp. 29 659–29 668, 2019. 1

    work page 2019

  14. [14]

    Toward automatic mask and source optimization for optical lithography,

    A. Erdmann, T. Fuehner, T. Schnattinger, and B. Tollkuehn, “Toward automatic mask and source optimization for optical lithography,” inOptical Microlithography XVII, vol. 5377. International Society for Optics and Photonics, 2004, pp. 646–657. 1

    work page 2004

  15. [15]

    Global source optimisation based on adaptive nonlinear particle swarm optimisation al- gorithm for inverse lithography,

    H. Sun, J. Du, C. Jin, J. Feng, J. Wang, S. Hu, and J. Liu, “Global source optimisation based on adaptive nonlinear particle swarm optimisation al- gorithm for inverse lithography,” IEEE Photonics Journal , vol. 13, no. 4, pp. 1–7, 2021. 1

    work page 2021

  16. [16]

    GAN-OPC: Mask optimization with lithography-guided generative adversarial nets,

    H. Yang, S. Li, Z. Deng, Y. Ma, B. Yu, and E. F. Young, “GAN-OPC: Mask optimization with lithography-guided generative adversarial nets,” 24 0 50 100 150 iteration step 0 500 1000 1500 2000 2500error 2=0.015 2=0.02 Figure 10: The output pattern with different β2. IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, vol. 39, no. 10...

    work page 2000

  17. [17]

    Generic lithography mod- eling with dual-band optics-inspired neural networks,

    H. Yang, Z. Li, K. Sastry, S. Mukhopadhyay, M. Kilgard, A. Anand- kumar, B. Khailany, V. Singh, and H. Ren, “Generic lithography mod- eling with dual-band optics-inspired neural networks,” arXiv preprint arXiv:2203.08616, 2022. 1

    work page arXiv 2022

  18. [18]

    Inverse lithography technology: 30 years from concept to prac- tical, full-chip reality,

    L. Pang, “Inverse lithography technology: 30 years from concept to prac- tical, full-chip reality,” Journal of Micro/Nanopatterning, Materials, and Metrology, vol. 20, no. 3, p. 030901, 2021. 1

    work page 2021

  19. [19]

    H. W. Engl, M. Hanke, and A. Neubauer, Regularization of Inverse Prob- lems. Kluwer Academic Publishers, 2000. 1

    work page 2000

  20. [20]

    Mask design for optical microlithogra- phy—an inverse imaging problem,

    A. Poonawala and P. Milanfar, “Mask design for optical microlithogra- phy—an inverse imaging problem,” IEEE Transactions on Image Process- ing, vol. 16, no. 3, pp. 774–788, 2007. 1

    work page 2007

  21. [21]

    Nonlinear compressive inverse lithography aided by low-rank regularization,

    X. Ma, Z. Wang, J. Zhu, S. Zhang, G. R. Arce, and S. Zhao, “Nonlinear compressive inverse lithography aided by low-rank regularization,” Optics express, vol. 27, no. 21, pp. 29 992–30 008, 2019. 1, 2.3

    work page 2019

  22. [22]

    Edge-preserving and scale-dependent properties of total variation regularization,

    D. Strong and T. Chan, “Edge-preserving and scale-dependent properties of total variation regularization,” Inverse problems, vol. 19, no. 6, p. S165,

    work page

  23. [23]

    S. K. Choy, N. Jia, C. S. Tong, M. L. Tang, and E. Y. Lam, “A robust computational algorithm for inverse photomask synthesis in optical projec- 25 (a) 50 100 150 200 50 100 150 200 (b) 50 100 150 200 50 100 150 200 (c) 50 100 150 200 50 100 150 200 (d) 50 100 150 200 50 100 150 200 (e) 50 100 150 200 50 100 150 200 (f) 50 100 150 200 50 100 150 200 (g) 50...

    work page 2012

  24. [24]

    An augmented lagrangian method for identifying discontinuous parameters in elliptic systems,

    Z. Chen and J. Zou, “An augmented lagrangian method for identifying discontinuous parameters in elliptic systems,” SIAM Journal on Control and Optimization, vol. 37, no. 3, pp. 892–910, 1999. 1, 2.4, 2.4

    work page 1999

  25. [25]

    An algorithm for total variation minimization and appli- cations,

    A. Chambolle, “An algorithm for total variation minimization and appli- cations,” Journal of Mathematical imaging and vision , vol. 20, no. 1, pp. 89–97, 2004. 1

    work page 2004

  26. [26]

    A nonlinear primal-dual method for total variation-based image restoration,

    T. F. Chan, G. H. Golub, and P. Mulet, “A nonlinear primal-dual method for total variation-based image restoration,” SIAM journal on scientific computing, vol. 20, no. 6, pp. 1964–1977, 1999. 1

    work page 1964

  27. [27]

    An iterative regu- larization method for total variation-based image restoration,

    S. Osher, M. Burger, D. Goldfarb, J. Xu, and W. Yin, “An iterative regu- larization method for total variation-based image restoration,” Multiscale Modeling & Simulation , vol. 4, no. 2, pp. 460–489, 2005. 1 26 (a) 50 100 150 200 250 50 100 150 200 250 (b) 50 100 150 200 250 50 100 150 200 250 (c) 50 100 150 200 250 50 100 150 200 250 (d) 50 100 150 200 2...

    work page 2005

  28. [28]

    A dual algorithm for the solution of nonlin- ear variational problems via finite element approximation,

    D. Gabay and B. Mercier, “A dual algorithm for the solution of nonlin- ear variational problems via finite element approximation,” Computers & mathematics with applications , vol. 2, no. 1, pp. 17–40, 1976. 1

    work page 1976

  29. [29]

    Alternating direction methods for classical and ptychographic phase retrieval,

    Z. Wen, C. Yang, X. Liu, and S. Marchesini, “Alternating direction methods for classical and ptychographic phase retrieval,” Inverse Problems, vol. 28, no. 11, p. 115010, 2012. 1

    work page 2012

  30. [30]

    A nonconvex ADMM algorithm for group sparsity with sparse groups,

    R. Chartrand and B. Wohlberg, “A nonconvex ADMM algorithm for group sparsity with sparse groups,” in 2013 IEEE international conference on acoustics, speech and signal processing. IEEE, 2013, pp. 6009–6013. 1

    work page 2013

  31. [31]

    A splitting method for orthogonality constrained problems,

    R. Lai and S. Osher, “A splitting method for orthogonality constrained problems,” Journal of Scientific Computing , vol. 58, no. 2, pp. 431–449,

    work page

  32. [32]

    1 27 (a) 50 100 150 200 250 50 100 150 200 250 (b) 50 100 150 200 250 50 100 150 200 250 (c) 50 100 150 200 250 50 100 150 200 250 (d) 50 100 150 200 250 50 100 150 200 250 (e) 50 100 150 200 250 50 100 150 200 250 (f) 50 100 150 200 250 50 100 150 200 250 (g) 50 100 150 200 250 50 100 150 200 250 (h) 50 100 150 200 250 50 100 150 200 250 (i) 50 100 150 2...

    work page

  33. [33]

    Global convergence of ADMM in non- convex nonsmooth optimization,

    Y. Wang, W. Yin, and J. Zeng, “Global convergence of ADMM in non- convex nonsmooth optimization,” Journal of Scientific Computing , vol. 78, no. 1, pp. 29–63, 2019. 1, 3.1

    work page 2019

  34. [34]

    SPIE press, 2001, vol

    Wong, Alfred Kwok-Kit, Resolution enhancement techniques in optical lithography. SPIE press, 2001, vol. 47. 2.1

    work page 2001

  35. [35]

    A. K.-K. Wong, Optical imaging in projection microlithography . SPIE press, 2005, vol. 66. 2.1, 2.3

    work page 2005

  36. [36]

    Ma and G

    X. Ma and G. R. Arce, Computational lithography. John Wiley & Sons,

    work page

  37. [37]

    Efficient source and mask optimization with augmented lagrangian methods in optical lithography,

    J. Li, S. Liu, and E. Y. Lam, “Efficient source and mask optimization with augmented lagrangian methods in optical lithography,”Optics express, vol. 21, no. 7, pp. 8076–8090, 2013. 2.3 28

    work page 2013

  38. [38]

    OPC and PSM design using inverse lithog- raphy: a nonlinear optimization approach,

    A. Poonawala and P. Milanfar, “OPC and PSM design using inverse lithog- raphy: a nonlinear optimization approach,” in Optical Microlithography XIX, vol. 6154. International Society for Optics and Photonics, 2006, p. 61543H. 2.3

    work page 2006

  39. [39]

    T. F. Chan and J. Shen, Image processing and analysis: variational, PDE, wavelet, and stochastic methods. Society for Industrial and Applied Mathematics, 2005. [Online]. Available: https://portal.igpublish.com/ iglibrary/search/SIAMB0000033.html 2.3

    work page 2005

  40. [40]

    R. A. Adams and J. J. Fournier, Sobolev spaces. Elsevier, 2003. 2.4, 2.4

    work page 2003

  41. [41]

    Nesterov et al., Lectures on convex optimization

    Y. Nesterov et al., Lectures on convex optimization . Springer, 2018, vol

    work page 2018

  42. [42]

    The split bregman method for L1-regularized problems,

    T. Goldstein and S. Osher, “The split bregman method for L1-regularized problems,” SIAM journal on imaging sciences , vol. 2, no. 2, pp. 323–343,

    work page

  43. [43]

    Numerical optimization,

    N. Jorge and J. W. Stephen, “Numerical optimization,” 2006. 3.2.1 29

    work page 2006