A Damped Subspace Splitting Algorithm for Constrained Density Functional Theory

Guanghui Hu; Xin Liu; Yuanming Su; Yukuan Hu

REVIEW 3 major objections 4 minor 69 references

A Damped Subspace Splitting Algorithm for Constrained Density Functional Theory

T0 review · 3 major / 4 minor · reviewed 2026-08-08 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Constrained DFT can be solved by a single-loop damped ADMM whose every accumulation point is a $\delta$-approximate KKT point, giving the field its first rigorous convergence guarantee.

desk verdict A serious, honest optimization paper that gives the first convergence guarantee for a damped subspace-splitting ADMM for CDFT, under an unverified eigengap assumption and for a parameter setting the authors do not use in their own benchmarks. read the letter →

arxiv 2608.05682 v1 pith:IVSUEEOC submitted 2026-08-06 math.OC physics.chem-phphysics.comp-ph

Yuanming Su , Yukuan Hu , Xin Liu , Guanghui Hu This is my paper · ORCID

classification math.OCphysics.chem-phphysics.comp-ph MSC 90C2690C3065K0565F15

keywords constraineddensityfunctionaltheoryCDFTalternatingdirectionmethodofmultiplierssubspacesplittingStiefelmanifoldnonconvexquadraticconstraintsdampeddualascentapproximateKKTpoints

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

Constrained density functional theory (CDFT) is the standard way to describe charge-localized and excited electronic states, but its discretized optimization problem couples an orthogonality constraint on orbitals with nonconvex quadratic density constraints, and existing solvers either violate the constraints or rely on fragile nested iterations. The paper shows that the two groups of constraints can be decoupled exactly: by exploiting rotation invariance, the density matrix $XX^\top$ is treated as a separate variable $Y$ constrained only by linear trace conditions, linked back to $X$ by $Y=XX^\top$. On this reformulation the paper builds DASSP, a single-loop damped alternating direction method of multipliers, and proves that under natural smoothness and eigengap assumptions every accumulation point is a $\delta$-approximate KKT point whose stationarity and feasibility errors are bounded by constants times $\delta$. A sympathetic reader would care because this is the first CDFT solver with a rigorous convergence guarantee, and the accuracy versus speed trade-off is controlled by one damping parameter.

What carries the argument

The load-bearing object is the subspace-splitting reformulation $Y=XX^\top$ combined with the damped augmented Lagrangian $\mathcal{L}_{\beta,\delta}(X,Y,\Lambda)=f(Y)+\langle(1-\delta)\Lambda,\,Y-XX^\top\rangle+(\beta/2)\|Y-XX^\top\|_F^2$. The affine set $\mathcal{Y}=\{Y:\,\mathrm{Tr}(W_jY)=b_j\}$ makes the nonconvex quadratic constraints linear, and the $X$-subproblem becomes a spectral step: compute the dominant $p$-dimensional invariant subspace of $A^{(k)}=Y^{(k)}+(1-\delta)/\beta\,\Lambda^{(k)}$. The $Y$-subproblem is a single projected gradient step onto $\mathcal{Y}$, with a closed-form projection requiring only the solution of the $m\times m$ normal system $M\hat\mu=-c$, and the dual variable is updated by the damped ascent $\Lambda^{(k+1)}=(1-\delta)\Lambda^{(k)}+\tau\beta(Y^{(k+1)}-X^{(k+1)}X^{(k+1)\top})$. The convergence proof is carried by the surrogate sequence $\Psi_k$, which adds a scaled dual-difference penalty to the damped augmented Lagrangian; Lemmas 3.5 and 3.6 show $\Psi_k$ descends each iteration and is bounded below, yielding the vanishing increments and the $\delta$-proportional stationarity and feasibility bounds.

What would settle it

Engineer a CDFT instance where the $p$-th and $(p+1)$-th eigenvalues of $A^{(k)}$ become equal at some iteration (for example, two identical, widely separated fragments with a degeneracy in the density-matrix subspace) and check whether the iterates still converge to a $\delta$-approximate KKT point; if they do not, Theorem 3.1 is false as stated. A complementary check is to measure the residuals at the accumulation point for $\delta=10^{-2}$, $10^{-4}$, and $10^{-6}$ and verify they decrease proportionally to $\delta$, as the theorem's bounds predict.

Watch

Extended reading notes

Core claim

The central claim is that DASSP, a damped ADMM for the subspace-splitting reformulation of CDFT, converges to approximately KKT points with an error that the user can preset. Theorem 3.1 states that if Assumption 3.1 holds, every accumulation point $(X^\star,Y^\star,\Lambda^\star)$ admits multipliers $\Sigma^\star$, $\mu^\star$ such that the stationarity residual $\|F(X^\star,\mu^\star)X^\star-X^\star\Sigma^\star\|_F$ is at most $L\sqrt{p(n-p)}/(\tau\beta)\|\Lambda^\star\|_F\,\delta$ and each quadratic constraint violation $|\mathrm{Tr}(X^{\star\top}W_jX^\star)-b_j|$ is at most $(1/(\tau\beta))\|\Lambda^\star\|_F\|W_j\|_F\,\delta$. The convergence mechanism is a surrogate-sequence argument: a constructed sequence $\Psi_k$ decreases at every iteration and is bounded below, which forces the primal and dual increments to vanish and controls the dual iterates, the usual obstacle for nonconvex ADMM. The feasibility error comes from the limiting alignment residual $R^\star=Y^\star-X^\star X^{\star\top}$; because $Y^\star$ is exactly feasible, the residual is absorbed by the dual update as $\delta/(\tau\beta)\|\Lambda^\star\|_F$. The paper also reports numerical evidence that DASSP reaches feasibility violations near $10^{-9}$ while the double-loop baseline struggles on systems with tiny eigengaps.

Load-bearing premise

At every step, the matrix whose top $p$ eigenvectors define the new orbitals must keep its $p$-th and $(p+1)$-th eigenvalues separated by a fixed positive gap; if the gap closes at any iteration, the proof's central descent estimate fails.

Editorial extensions

If this is right

CDFT calculations can be run as a single loop with no inner Newton iterations, so per-iteration cost is one eigensolve, one projected gradient step, and one closed-form dual update, all of standard numerical linear algebra.
The final accuracy is set in advance: choosing a smaller damping $\delta$ produces limit points that are $\delta$-proportionally closer to being stationary and feasible, at the cost of slower dual movement.
Penalty-based methods no longer need to be pushed to very large penalty parameters to force feasibility, so the conditioning problems documented for quadratic penalty SCF can be avoided.
The convergence theorem applies to any smooth objective on the Stiefel manifold with finitely many quadratic constraints satisfying the assumptions, not just to the Kohn-Sham energy; the paper's claims are formulated at that general level.
The undamped limit $\delta=0$, which the numerical experiments find fast and stable, is not covered by the theorem and is explicitly left open by the authors.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

A continuation schedule on $\delta$ (large damping for stability, then shrinking $\delta$ to refine accuracy) is a natural practical extension that the paper does not test; the $\delta$-proportional error bounds make the accuracy at each stage predictable.
Because DASSP does not rely on the effective Hamiltonian's eigengap but on the eigengap of $A^{(k)}$, it may be stable in charge-localization regimes where standard SCF-based CDFT oscillates; the water dimer cation experiment supports this, and a broader comparison on magnetic-moment constraints would test it.
The same subspace-splitting reformulation should extend to noncollinear CDFT with local magnetic-moment constraints, since those constraints are also quadratic in the orbitals; the paper lists this as future work, and the optimization proof is constraint-agnostic apart from the linear independence assumption.
The $\delta$-approximate KKT certificate could be turned into a practical stopping rule: stop when increments are small and report $\|\Lambda\|_F/(\tau\beta)\delta$ as a certified bound, though the paper does not propose such a protocol.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Quick take: this is a strong, honest optimization paper, but the headline guarantee is more conditional than the abstract suggests. The subspace-splitting reformulation is genuinely new for CDFT, and the damped ADMM analysis with a carefully constructed surrogate sequence is a real step for nonconvex, nonlinearly constrained problems. The numerical work is broad and supports the practical claims: DASSP hits feasibility around 1e-9 and beats the double-loop WV method in wall-clock time, especially when the effective Hamiltonian has a tiny eigengap. The water dimer case is convincing.

Soft spots, in order of importance. First, Assumption 3.1(A2) — uniform positive eigengap for A(k) — is load-bearing: Lemma 3.1, the only descent control on the X-block, uses it directly. The paper gives one numerical check on water dimer and a local-plausibility remark, but does not verify the eigengap on the charge-transfer benchmarks where Λ(k) moves a lot. If A(k) degenerates at some iteration, the proof stops. The stress-test is right about this, and the paper's own Section 4.4 shows how close CDFT systems can get to degeneracy. Second, the recommended default δ=0 is outside the theorem's δ>0 regime. The authors are explicit about this, and the δ=10^-10 curves are practically identical to δ=0, so it is a minor-to-moderate gap rather than a fatal one, but it does mean the benchmarked algorithm is not the one with the rigorous guarantee. Third, the guarantee is δ-approximate KKT; that is a reasonable price for damping, not a flaw.

On the citation pattern: Lemma 3.1 borrows from [45], which one of the authors co-authored, but that result is published and checkable; no circularity burden. No code is released, which slows reproducibility but is not a scientific defect.

Overall: a solid paper, honestly written, with a clear new contribution. The main claim should be softened in the abstract — "first damped ADMM variant for CDFT with rigorous guarantees under an eigengap assumption" rather than the unqualified version. I would send it to a serious referee on optimization and electronic structure; the referee should push for a sharper treatment of Assumption 3.1(A2) or a verification on the actual test problems, and ideally an analysis covering δ=0 or a change to the default parameter.

Referee Report

3 major / 4 minor

Summary. The paper reformulates the discretized constrained density functional theory problem (1.2) by introducing the projector variable Y = XX^T, which turns the nonconvex quadratic constraints into linear constraints in Y. A single-loop damped ADMM (DASSP) is proposed: an X-step that computes a dominant p-dimensional eigenspace of A(k) = Y(k) + (1 - δ)/β Λ(k), a Y-step that performs one projected gradient step on the affine set Y, and a damped dual update for Λ. The main convergence theorem (Theorem 3.1) states that under Assumption 3.1, including a uniform positive eigengap of A(k) and δ ∈ (0,1), every accumulation point is a δ-approximate KKT point with explicit stationarity and feasibility error bounds. Numerical experiments compare DASSP with quadratic-penalty and double-loop methods on charge-transfer systems and on a water-dimer cation charge-localization test, reporting high feasibility accuracy and speedups.

Significance. If the theorem is accepted, the paper delivers a genuinely single-loop algorithm with a nontrivial convergence guarantee for a nonconvex problem that has lacked such guarantees; the surrogate-sequence proof is explicit, the descent constants are derived from the stated parameter assumptions, and the numerical study is broader than typical for an optimization methods paper. The paper is also honest about the scope of the theory, explicitly flagging the δ = 0 limitation. The main gap is that the central guarantee is conditional on an eigengap assumption that is not verified for the main benchmarks, and the recommended experimental setting (δ = 0, adaptive stepsize, τ = 0.2) is outside the theorem's parameter regime; therefore the advertised 'first rigorous convergence guarantee' is not yet demonstrated for the algorithm as actually used.

major comments (3)

[§4.1 and Theorem 3.1] The convergence theorem is proved only for δ ∈ (0,1): Assumption 3.1(A5) contains δ in the numerator and requires c > 2(1−δ)(2−δ)/δ, so the case δ = 0 is excluded. Section 4.1 nevertheless adopts δ = 0 as the default for all subsequent experiments, with the sentence 'even though its convergence is not covered in this work.' The abstract and contribution statements should therefore be scoped to the damped variant; as written, the claim that DASSP is the first algorithm with rigorous convergence guarantees for CDFT overstates what is proved. Please either extend the analysis to δ = 0 or restrict the claim and describe the undamped/default version as a heuristic whose convergence is only demonstrated numerically.
[Assumption 3.1(A2), Lemma 3.1, §4.4] The uniform eigengap γ_k ≥ γ > 0 for A(k) is load-bearing: Lemma 3.1 uses it to turn the X-update into a descent step, and without it the telescoping argument for the surrogate sequence Ψ_k has no control over the X-block. The paper does not derive this condition from problem data; Remark 3.1 only argues local plausibility near a fixed point with large β. The only direct numerical evidence is Section 4.4 for the water-dimer cation with δ = 10^-10, and the charge-transfer benchmarks in Tables 1 and 2 do not report γ_k. Given that the paper itself finds effective-Hamiltonian eigengaps as small as ~10^-12 in a CDFT problem, a vanishing gap for A(k) is a live risk in the tested applications. Please either prove a sufficient condition for A2 from the problem structure, or monitor γ_k for every reported benchmark and state the convergence claim as conditional on this verified condition.
[Assumption 3.1(A4)-(A5) vs §4.1 parameters] The certified parameter regime is much narrower than the experimental one. Assumption 3.1(A4) requires a constant stepsize η ∈ (0, 2/(β+L)), while the default η_k = ηABB is a clipped adaptive Barzilai-Borwein stepsize. Moreover, for the value δ = 10^-10 used in the δ-sweep and in Section 4.4, Assumption 3.1(A5) forces τ to be of order δ/(2c) ≈ 10^-21 (with c ≈ 2(1−δ)(2−δ)/δ), whereas the experiments use τ = 0.2. Thus even the damped experiments do not satisfy the theorem's parameter conditions. The remark in Assumption 3.1(A4) that variable stepsizes can be handled by a 'straightforward but tedious' extension is not a proof. Please either prove convergence for the adaptive, under-relaxed parameter choices actually used, or report experiments in the certified regime (constant η, sufficiently small τ) and confirm that the qualitative conclusions are unchanged.

minor comments (4)

[Eq. (2.7)-(2.8)] The invertibility of the Gram matrix M is attributed to Assumption 1.1(A2); since A2 is stated for the matrices P⊥_X W_j X at local minimizers, not for the W_j themselves, a one-sentence justification outside the proof would help the reader.
[§4.4, Figure 8] The text says the initial eigengap of A(0) may vanish and the figure reports a minimal value of 1.110e-16; please clarify whether this value occurs at k = 0 and whether the plateau value is reached before the iteration count reported, so the reader can judge how representative the eigengap history is.
[Figures 6-7] Several axis labels and legends in Figures 6 and 7 contain garbled text ('le el', 'Oute SCF ite ation', 'Subspace di tance') that should be corrected in a revised version.
[§2, Y-update] The set Y defined in (2.2) is an affine subspace, not a manifold; calling TY its 'tangent space' is nonstandard and could be replaced by 'parallel subspace' or a similar term.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the convergence theorem is derived from stated assumptions, and the only self-citation is to a published, checkable spectral lemma that does not supply the theorem's conclusion.

full rationale

The central claim, Theorem 3.1, is proved within the paper from Assumption 3.1 via explicit descent inequalities (Lemmas 3.1-3.7 and Corollary 3.1); no fitted parameter or empirical value is used in the proof. The subspace-splitting reformulation (2.1) is an exact equivalent problem by construction, not a prediction derived from its own output. Lemma 3.1 invokes [45, Lemma 3.1], a published spectral perturbation bound, and this is independent support rather than circularity because it is a parameter-free, externally checkable statement whose assumptions do not include the target convergence result. Assumption 3.1(A2), the uniform eigengap, is an explicitly stated assumption; the paper acknowledges it cannot remove it and provides only numerical evidence in Section 4.4, which is a limitation, not a circular step. Likewise, the parameter tuning in Section 4.1 is empirical and does not enter the theory, and the default choice delta = 0 being outside the proved damped regime is a gap between theory and practice, not a reduction of the theorem to its inputs. The derived KKT error bounds are genuine consequences of the damping construction, with constants expressed in terms of the accumulated multiplier Lambda*, rather than being imposed by definition. Overall, no load-bearing step reduces to its own input by construction or to an unverified self-citation.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The theoretical claim rests entirely on standard smoothness and constraint-qualification assumptions, a uniform eigengap that is numerically supported but not proven for CDFT, and explicit parameter conditions. No physical constants are fitted, and the convergence proof contains no fitted parameters. The algorithm's four hyperparameters are tuned for performance in Section 4.1, which is standard engineering practice and does not enter the theorem. The Y variable is a reformulation device, not a new physical entity.

free parameters (4)

β (penalty parameter) = 20 (default)
Tuned on synthetic CO2/C2H6/C6H5OCH3 instances in Section 4.1; theory only requires β>2(L+1).
τ (under-relaxation) = 0.2 (default)
Tuned in Section 4.1; theory requires τ < δ/(2c) min(...) for δ>0.
δ (damping parameter) = 0 (default in experiments)
Tuned in Section 4.1; theory requires δ>0. The paper uses δ=0 in practice, which is outside the convergence theorem.
η_k stepsize strategy (clipped ABB) = η_ABB, clipped to [1e-6,10]
Chosen by comparison with constant and Armijo strategies in Section 4.1; theory requires constant η in (0, 2/(β+L)).

assumptions (4)

domain assumption Feasibility and LICQ (Assumption 1.1)
The problem has at least one feasible point and the Riemannian LICQ holds at local minimizers, needed for KKT conditions.
domain assumption f is L-smooth (Assumption 3.1 A1)
The Kohn-Sham energy restricted to the considered XC functionals has Lipschitz gradient; used in descent lemmas.
ad hoc to paper Uniform eigengap of A(k) (Assumption 3.1 A2)
Requires λ_p(A(k))-λ_{p+1}(A(k))≥γ>0 for all k; used to prove Lemma 3.1. Not proven for CDFT; only numerically supported in Section 4.4.
standard math Parameter bounds (Assumptions 3.1 A3-A5)
β>2(L+1), η∈(0,2/(β+L)), τ small enough; these are algebraic conditions ensuring the surrogate constants are positive.

how reviews work

0 comments

Cite this review

Pith. "Pith review of A Damped Subspace Splitting Algorithm for Constrained Density Functional Theory." pith.science (2026). https://pith.science/paper/IVSUEEOC

@misc{pith2026260805682,
  author       = {Pith},
  title        = {Pith review of: A Damped Subspace Splitting Algorithm for Constrained Density Functional Theory},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/IVSUEEOC}},
  note         = {Machine review of arXiv:2608.05682}
}

read the original abstract

Constrained density functional theory (CDFT) provides a powerful framework for describing electronically excited and charge-localized states, which underlie a broad range of physical and chemical phenomena. However, the discretized optimization problems arising from CDFT calculations remain challenging, owing to the presence of both the Stiefel manifold constraint and additional nonconvex quadratic constraints. Existing algorithms either fail to enforce the quadratic constraints with high accuracy or face convergence issues due to double-loop iterative structures. In this paper, we first derive a subspace-splitting reformulation that decouples the two groups of constraints, by exploiting the inherent rotation invariance and introducing a nonlinear subspace alignment constraint. Based on this reformulation, we propose a single-loop damped alternating direction method of multipliers, called DASSP. To the best of our knowledge, DASSP is the first algorithm for CDFT calculations with rigorous convergence guarantees. Each iteration of DASSP comprises a spectral minimization step, a projected gradient step, and a damped dual ascent step, all of which admit efficient implementations. Numerical results on synthetic and realistic CDFT problems demonstrate that DASSP attains high feasibility accuracy and exhibits favorable efficiency without compromising robustness. We expect that this work will pave the way toward reliable and efficient large-scale CDFT applications.

Figures

Figures reproduced from arXiv: 2608.05682 by the authors.

**Figure 2.** Comparison between different values of δ. The blue curves with circle markers, orange curves with square markers, green curves with triangle markers, red curves with diamond markers, purple curves with down-pointing triangle markers, and brown curves with star markers represent the results using δ = 10−2 , 10−4 , 10−6 , 10−8 , 10−10 , 0, respectively. From top to bottom: CO2, C2H6, and C6H5OCH3. From left to right: … view at source ↗

**Figure 3.** Comparison between different values of β. The blue curves with circle markers, orange curves with square markers, green curves with triangle markers, red curves with diamond markers, purple curves with down-pointing triangle markers, brown curves with star markers, and pink curves with pentagon markers represent the results using β = 5, 10, 20, 30, 40, 50, 100, respectively. From top to bottom: CO2, C2H6, and C6H5O… view at source ↗

Figures from the paper (5 more)

**Figure 4.** Figure 4: Comparison between different values of τ . The blue curves with circle markers, orange curves with square markers, green curves with triangle markers, red curves with diamond markers, purple curves with down-pointing triangle markers, brown curves with star markers, pi…

**Figure 5.** Figure 5: Energy curves of the QP-SCF with different values of [PITH_FULL_IMAGE:figures/full_fig_p019_5.png]

**Figure 6.** Figure 6: Success rates of DASSP and WV methods as functions o [PITH_FULL_IMAGE:figures/full_fig_p021_6.png]

**Figure 7.** Figure 7: Evolutions of the eigengap of the effective Hamilton [PITH_FULL_IMAGE:figures/full_fig_p022_7.png]

**Figure 8.** Figure 8: Evolutions of the eigengap of the matrix [PITH_FULL_IMAGE:figures/full_fig_p023_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

69 extracted references · 68 canonical work pages

[45]

X. Liu, X. Wang, Z. Wen, and Y. Yuan. On the convergence of the self-consistent ﬁeld iteration in Kohn-Sham density functional theory. SIAM J. Matrix Anal. Appl. , 35(2):546– 558, 2014

work page 2014
[1]

Absil, R

P.-A. Absil, R. Mahony, and R. Sepulchre. Optimization Algorithms on Matrix Manifolds . Princeton University Press, 2008

work page 2008
[2]

C. S. Ahart, K. M. Rosso, and J. Blumberger. Implementati on and validation of con- strained density functional theory forces in the CP2K packa ge. J. Chem. Theory Comput. , 18(7):4438–4446, 2022

work page 2022
[3]

Alml¨ of, K

J. Alml¨ of, K. Fægri Jr, and K. Korsell. Principles for a d irect SCF approach to LICAO- MOab-initio calculations. J. Comput. Chem. , 3(3):385–399, 1982

work page 1982
[4]

Bai, R.-C

Z. Bai, R.-C. Li, and D. Lu. Sharp estimation of convergen ce rate for self-consistent ﬁeld iteration to solve eigenvector-dependent nonlinear eigen value problems. SIAM J. Matrix Anal. Appl. , 43(1):301–327, 2022

work page 2022
[5]

Balzani, P

V. Balzani, P. Ceroni, and A. Juris. Photochemistry and Photophysics: Concepts, Research, Applications. John Wiley & Sons, 2014

work page 2014
[6]

Bergmann and R

R. Bergmann and R. Herzog. Intrinsic formulation of KKT c onditions and constraint qualiﬁcations on smooth manifolds. SIAM J. Optim. , 29(4):2423–2444, 2019

work page 2019
[7]

Bolte, S

J. Bolte, S. Sabach, and M. Teboulle. Nonconvex Lagrangi an-based optimization: monitor- ing schemes and global convergence. Math. Oper. Res. , 43(4):1210–1232, 2018

work page 2018

Show all 69 references

[8]

S. Boyd, N. Parikh, E. Chu, B. Peleato, and J. Eckstein. Di stributed optimization and statistical learning via the alternating direction method of multipliers. Found. Trends Mach. Learn., 3(1):1–122, 2011

2011
[9]

K. Burke. Perspective on density functional theory. J. Chem. Phys. , 136(15), 2012

2012
[10]

Z. Cai, K. Wang, Y. Xu, S.-H. Wei, and B. Xu. A self-adapti ve ﬁrst-principles approach for magnetic excited states. Quantum Front., 2(1):21, 2023

2023
[11]

Canc` es and G

E. Canc` es and G. Friesecke. Density Functional Theory: Modeling, Mathematical Analysi s, Computational Methods, and Applications . Mathematics and Molecular Modeling. Springer Cham, 2023

2023
[12]

Cohen, N

E. Cohen, N. Hallak, and M. Teboulle. A dynamic alternat ing direction of multipliers for nonconvex minimization with nonlinear functional equalit y constraints. J. Optim. Theory Appl., 193(1):324–353, 2022. 26

2022
[13]

Dai and R

Y.-H. Dai and R. Fletcher. Projected Barzilai-Borwein methods for large-scale box- constrained quadratic programming. Numer. Math. , 100(1):21–47, 2005

2005
[14]

P. H. Dederichs, S. Bl¨ ugel, R. Zeller, and H. Akai. Grou nd states of constrained systems: application to cerium impurities. Phys. Rev. Lett. , 53(26):2512, 1984

1984
[15]

K. Deng, J. Jin, J. Hu, and H. Wang. Adaptive Riemannian A DMM for nonsmooth op- timization: optimal complexity without smoothing. In The 39th Conference on Neural Information Processing Systems , 2025

2025
[16]

El Bourkhissi and I

L. El Bourkhissi and I. Necoara. Convergence rates for a n inexact linearized ADMM for nonsmooth nonconvex optimization with nonlinear equality constraints. Comput. Optim. Appl., pages 1–39, 2025

2025
[17]

Gabay and B

D. Gabay and B. Mercier. A dual algorithm for the solutio n of nonlinear variational prob- lems via ﬁnite element approximation. Comput. Math. Appl. , 2(1):17–40, 1976

1976
[18]

B. Gao, G. Hu, Y. Kuang, and X. Liu. An orthogonalization -free parallelizable framework for all-electron calculations in density functional theor y. SIAM J. Sci. Comput. , 44(3):B723– B745, 2022

2022
[19]

B. Gao, X. Liu, and Y.-X. Yuan. Parallelizable algorith ms for optimization problems with orthogonality constraints. SIAM J. Sci. Comput. , 41(3):A1949–A1983, 2019

2019
[20]

Garc ´ ıa, N

A. Garc ´ ıa, N. Papior, A. Akhtar, E. Artacho, V. Blum, E. Bosoni, P. Brandimarte, M. Brandbyge, J. I. Cerd´ a, F. Corsetti, R. Cuadrado, V. Dikan, J. Ferrer, J. Gale, P. Garc ´ ıa- Fern´ andez, V. M. Garc ´ ıa-Su´ arez, S. Garc ´ ıa, G. Huhs, S. Illera, R. Koryt´ ar, P. Kova...

2020
[21]

Glowinski and A

R. Glowinski and A. Marroco. Sur l’approximation, par ´ el´ ements ﬁnis d’ordre un, et la r´ esolution, par p´ enalisation-dualit´ e d’une classe de probl` emes de Dirichlet non lin´ eaires. RAIRO. Anal. num´ er., 9(R2):41–76, 1975

1975
[22]

G. H. Golub and C. F. Van Loan. Matrix Computations . Johns Hopkins University Press, 4th edition, 2013

2013
[23]

Gonze, B

X. Gonze, B. Seddon, J. A. Elliott, C. Tantardini, and A. V. Shapeev. Constrained density functional theory: a potential-based self-consistency ap proach. J. Chem. Theory Comput. , 18(10):6099–6110, 2022

2022
[24]

Grimme, J

S. Grimme, J. Antony, S. Ehrlich, and H. Krieg. A consist ent and accurate ab initio parametrization of density functional dispersion correct ion (DFT-D) for the 94 elements H-Pu. J. Chem. Phys. , 132(15), 2010

2010
[25]

Hajinezhad and M

D. Hajinezhad and M. Hong. Perturbed proximal primal-d ual algorithm for nonconvex nonsmooth optimization. Math. Program., 176(1):207–245, 2019

2019
[26]

Hallak and M

N. Hallak and M. Teboulle. An adaptive Lagrangian-base d scheme for nonconvex composite optimization. Math. Oper. Res. , 48(4):2337–2352, 2023

2023
[27]

D. Han. A survey on some recent developments of alternat ing direction method of multi- pliers. J. Oper. Res. Soc. China , 10(1):1–52, 2022. 27

2022
[28]

W. J. Hehre, R. F. Stewart, and J. A. Pople. Self-consist ent molecular-orbital methods. I. Use of Gaussian expansions of Slater-type atomic orbitals. J. Chem. Phys. , 51(6):2657–2664, 1969

1969
[29]

M. R. Hestenes. Multiplier and gradient methods. J. Optim. Theory Appl. , 4(5):303–320, 1969

1969
[30]

L. T. K. Hien and D. Papadimitriou. An inertial ADMM for a class of nonconvex composite optimization with nonlinear coupling constraints. J. Global Optim. , 89(4):927–948, 2024

2024
[31]

Hohenberg and W

P. Hohenberg and W. Kohn. Inhomogeneous electron gas. Phys. Rev., 136(3B):B864, 1964

1964
[32]

J. Hu, X. Liu, Z. Wen, and Y.-X. Yuan. A brief introductio n to manifold optimization. J. Oper. Res. Soc. China , 8(2):199–248, 2020

2020
[33]

R. D. Johnson III. NIST Computational Chemistry Compar ison and Benchmark Database. NIST Standard Reference Database Number 101, 2022

2022
[34]

Kaduk, T

B. Kaduk, T. Kowalczyk, and T. Van Voorhis. Constrained density functional theory. Chem. Rev. , 112(1):321–370, 2012

2012
[35]

Kohn and L

W. Kohn and L. J. Sham. Self-consistent equations inclu ding exchange and correlation eﬀects. Phys. Rev., 140(4A):A1133, 1965

1965
[36]

Kong and R

W. Kong and R. D. C. Monteiro. An accelerated inexact dam pened augmented Lagrangian method for linearly-constrained nonconvex composite opti mization problems. Comput. Op- tim. Appl. , 85(2):509–545, 2023

2023
[37]

Kong and R

W. Kong and R. D. C. Monteiro. Global complexity bound of a proximal ADMM for linearly constrained nonseparable nonconvex composite pr ogramming. SIAM J. Optim. , 34(1):201–224, 2024

2024
[38]

Kovnatsky, K

A. Kovnatsky, K. Glashoﬀ, and M. M. Bronstein. MADMM: a ge neric algorithm for non- smooth optimization on manifolds. In B. Leibe, J. Matas, N. S ebe, and M. Welling, editors, European Conference on Computer Vision , volume 9909 of Lecture Notes in Computer Science, pages 680–...

2016
[39]

Kresse and J

G. Kresse and J. Furthm¨ uller. Eﬃciency of ab-initio to tal energy calculations for metals and semiconductors using a plane-wave basis set. Comput. Mater. Sci. , 6(1):15–50, 1996

1996
[40]

Kresse and J

G. Kresse and J. Furthm¨ uller. Eﬃcient iterative schem es for ab initio total-energy calcula- tions using a plane-wave basis set. Phys. Rev. B , 54(16):11169, 1996

1996
[41]

T. D. K¨ uhne, M. Iannuzzi, M. Del Ben, V. V. Rybkin, P. See wald, F. Stein, T. Laino, R. Z. Khaliullin, O. Sch¨ utt, F. Schiﬀmann, D. Golze, J. Wilhe lm, S. Chulkov, M. H. Bani- Hashemian, V. Weber, U. Borˇ stnik, M. Taillefumier, A. S. Jakobovits, A. Lazzaro, H. Pabst, T. M¨...

2020
[42]

Lai and S

R. Lai and S. Osher. A splitting method for orthogonalit y constrained problems. J. Sci. Comput., 58(2):431–449, 2014. 28

2014
[43]

J. Li, S. Ma, and T. Srivastava. A Riemannian alternatin g direction method of multipliers. Math. Oper. Res. , 50(4):3222–3242, 2025

2025
[44]

L. Lin, J. Lu, and L. Ying. Numerical methods for Kohn-Sh am density functional theory. Acta Numer. , 28:405–539, 2019

2019
[46]

X. Liu, Z. Wen, X. Wang, M. Ulbrich, and Y. Yuan. On the ana lysis of the discretized Kohn-Sham density functional theory. SIAM J. Numer. Anal. , 53(4):1758–1785, 2015

2015
[47]

C. Lu, J. Feng, Z. Lin, and S. Yan. Nonconvex sparse spect ral clustering by alternating direction method of multipliers and its convergence analys is. In Proceedings of the AAAI Conference on Artiﬁcial Intelligence , volume 32, 2018

2018
[48]

Ma and S

P.-W. Ma and S. L. Dudarev. Constrained density functio nal for noncollinear magnetism. Phys. Rev. B , 91(5):054420, 2015

2015
[49]

J. G. Melo, R. D. C. Monteiro, and H. Wang. Iteration-Com plexity of an Inexact Proxi- mal Accelerated Augmented Lagrangian Method for Solving Li nearly Constrained Smooth Nonconvex Composite Optimization Problems. arXiv preprin t arXiv:2006.08048, 2020

2006 arXiv
[50]

J. P. Perdew, K. Burke, and M. Ernzerhof. Generalized gr adient approximation made simple. Phys. Rev. Lett. , 77(18):3865–3868, 1996

1996
[51]

M. J. D. Powell. A method for nonlinear constraints in mi nimization problems. In R. Fletcher, editor, Optimization, pages 283–298. Academic Press, 1969

1969
[52]

P. Pulay. Convergence acceleration of iterative seque nces. The case of SCF iteration. Chem. Phys. Lett., 73(2):393–398, 1980

1980
[53]

P. Pulay. Improved SCF convergence acceleration. J. Comput. Chem. , 3(4):556–560, 1982

1982
[54]

Sun and X

K. Sun and X. A. Sun. Dual descent augmented Lagrangian m ethod and alternating direc- tion method of multipliers. SIAM J. Optim. , 34(2):1679–1707, 2024

2024
[55]

Q. Sun, X. Zhang, S. Banerjee, P. Bao, M. Barbry, N. S. Blu nt, N. A. Bogdanov, G. H. Booth, J. Chen, Z.-H. Cui, J. J. Eriksen, Y. Gao, S. Guo, J. Her mann, M. R. Hermes, K. Koh, P. Koval, S. Lehtola, Z. Li, J. Liu, N. Mardirossian, J . D. McClain, M. Motta, B. Mussard, H. Q. ...

2020
[56]

N. J. Turro, V. Ramamurthy, and J. C. Scaiano. Principles of Molecular Photochemistry: an Introduction. University Science Books, 2009

2009
[57]

J. H. Van Lenthe, R. Zwaans, H. J. J. Van Dam, and M. F. Gues t. Starting SCF calculations by superposition of atomic densities. J. Comput. Chem. , 27(8):926–932, 2006

2006
[58]

L. Wang, X. Liu, and Y. Zhang. A communication-eﬃcient a nd privacy-aware distributed algorithm for sparse PCA. Comput. Optim. Appl. , 85(3):1033–1072, 2023. 29

2023
[59]

L. Wang, X. Liu, and Y. Zhang. Seeking consensus on subsp aces in federated principal component analysis. J. Optim. Theory Appl. , 203(1):529–561, 2024

2024
[60]

Weigend and R

F. Weigend and R. Ahlrichs. Balanced basis sets of split valence, triple zeta valence and quadruple zeta valence quality for H to Rn: design and assess ment of accuracy. Phys. Chem. Chem. Phys. , 7(18):3297–3305, 2005

2005
[61]

Wu and T

Q. Wu and T. Van Voorhis. Direct optimization method to s tudy constrained systems within density-functional theory. Phys. Rev. A , 72(2):024502, 2005

2005
[62]

Wu and T

Q. Wu and T. Van Voorhis. Constrained density functiona l theory and its application in long-range electron transfer. J. Chem. Theory Comput. , 2(3):765–774, 2006

2006
[63]

L. Yang, T. K. Pong, and X. Chen. Alternating direction m ethod of multipliers for a class of nonconvex and nonsmooth problems with application s to background/foreground extraction. SIAM J. Imag. Sci. , 10(1):74–110, 2017

2017
[64]

H. Yu, B. Liu, Y. Zhong, L. Hong, J. Ji, C. Xu, X. Gong, and H . Xiang. Physics-informed time-reversal equivariant neural network potential for ma gnetic materials. Phys. Rev. B , 110(10):104427, 2024

2024
[65]

H. Yu, Y. Zhong, L. Hong, C. Xu, W. Ren, X. Gong, and H. Xian g. Spin-dependent graph neural network potential for magnetic materials. Phys. Rev. B , 109(14):144426, 2024

2024
[66]

Zhang, S

J. Zhang, S. Ma, and S. Zhang. Primal-dual optimization algorithms over Riemannian manifolds: an iteration complexity analysis. Math. Program., 184(1):445–490, 2020

2020
[67]

Zheng, X

D. Zheng, X. Peng, Y. Huang, Y. Wang, D. Zhang, Z. Huang, Z . Cai, L. Zhang, M. Chen, B. Xu, and W. Zhou. Integrating deep-learning-based magnet ic model and non-collinear spin-constrained method: methodology, implementation an d application. npj Comput. Mater., 2026

2026
[68]

Y. Zhou, X. Shi, L. Guo, J. Cao, and M. Abdel-Aty. Perturb ed Proximal Gradient ADMM for Nonconvex Composite Optimization. arXiv preprint arXi v:2504.12759, 2025

2025
[69]

D. Zhu, L. Zhao, and S. Zhang. A ﬁrst-order primal-dual m ethod for nonconvex constrained optimization based on the augmented Lagrangian. Math. Oper. Res. , 49(1):125–150, 2024. 30

2024

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.

[1] [45]

X. Liu, X. Wang, Z. Wen, and Y. Yuan. On the convergence of the self-consistent ﬁeld iteration in Kohn-Sham density functional theory. SIAM J. Matrix Anal. Appl. , 35(2):546– 558, 2014

work page 2014

[2] [1]

Absil, R

P.-A. Absil, R. Mahony, and R. Sepulchre. Optimization Algorithms on Matrix Manifolds . Princeton University Press, 2008

work page 2008

[3] [2]

C. S. Ahart, K. M. Rosso, and J. Blumberger. Implementati on and validation of con- strained density functional theory forces in the CP2K packa ge. J. Chem. Theory Comput. , 18(7):4438–4446, 2022

work page 2022

[4] [3]

Alml¨ of, K

J. Alml¨ of, K. Fægri Jr, and K. Korsell. Principles for a d irect SCF approach to LICAO- MOab-initio calculations. J. Comput. Chem. , 3(3):385–399, 1982

work page 1982

[5] [4]

Bai, R.-C

Z. Bai, R.-C. Li, and D. Lu. Sharp estimation of convergen ce rate for self-consistent ﬁeld iteration to solve eigenvector-dependent nonlinear eigen value problems. SIAM J. Matrix Anal. Appl. , 43(1):301–327, 2022

work page 2022

[6] [5]

Balzani, P

V. Balzani, P. Ceroni, and A. Juris. Photochemistry and Photophysics: Concepts, Research, Applications. John Wiley & Sons, 2014

work page 2014

[7] [6]

Bergmann and R

R. Bergmann and R. Herzog. Intrinsic formulation of KKT c onditions and constraint qualiﬁcations on smooth manifolds. SIAM J. Optim. , 29(4):2423–2444, 2019

work page 2019

[8] [7]

Bolte, S

J. Bolte, S. Sabach, and M. Teboulle. Nonconvex Lagrangi an-based optimization: monitor- ing schemes and global convergence. Math. Oper. Res. , 43(4):1210–1232, 2018

work page 2018