pith. sign in

arxiv: 2605.16960 · v1 · pith:SNWO7HP3new · submitted 2026-05-16 · 💻 cs.IT · math.IT

Achieving α-Fairness in Clustered Cell-Free Networking: A Tight Relaxation Approach

Chaowen Deng , Jie Fan , Boxiang Ren , Ziyuan Lyu , Jingchen Peng , Hao Wu , Junyuan Wang This is my paper

Pith reviewed 2026-05-19 19:00 UTC · model grok-4.3

classification 💻 cs.IT math.IT
keywords clustered cell-free networkingalpha-fairnesscontinuous relaxationinteger programmingergodic capacitymassive MIMOspectral efficiencynetwork fairness
0
0 comments X

The pith

Reformulating the clustering problem as a continuous optimization establishes exact equivalence to the integer program for alpha-fairness across four cases.

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

Clustered cell-free networks combine high performance of cell-free massive MIMO with cellular scalability, yet fairness across subnetworks remains hard to achieve with discrete clustering decisions. The paper introduces a tunable alpha-fairness objective that trades total spectral efficiency against equitable rates among subnetworks. Substituting the ergodic sum capacity with its closed-form deterministic equivalent converts the combinatorial clustering task into a continuous optimization. The authors divide the problem into four cases by alpha value and prove that the continuous version is exactly equivalent to the original integer program in every case, which permits simple convergent algorithms. Simulations indicate the resulting clusterings raise fairness metrics substantially while reducing aggregate throughput only slightly.

Core claim

Using the closed-form deterministic equivalent of the ergodic sum capacity, the combinatorial clustering problem is reformulated as a continuous optimization problem. For each of four cases according to the value of alpha, the exact equivalence between the original integer program and its continuous relaxation is established, and efficient algorithms with guaranteed convergence are developed.

What carries the argument

The tight continuous relaxation of the integer alpha-fair clustering program, which becomes equivalent to the discrete version once the deterministic equivalent of capacity is substituted.

If this is right

  • The alpha parameter provides a unified way to tune the balance between aggregate spectral efficiency and inter-subnetwork fairness.
  • For each of the four alpha cases the relaxation is tight, so the continuous solution is optimal for the original integer problem.
  • Efficient algorithms with guaranteed convergence are obtained for all four cases.
  • The approach yields up to 11 percent higher Jain's fairness index and 45 percent higher minimum subnetwork capacity at the cost of only 5 percent lower aggregate throughput.

Where Pith is reading between the lines

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

  • The same tightness argument might apply to other discrete assignment problems such as user-to-subnetwork association or joint clustering and power control.
  • Direct validation of the deterministic equivalent on measured channels rather than ideal models would test how much the equivalence survives in practice.
  • The method could be paired with low-complexity heuristics for very large networks where even the relaxed problem remains expensive.

Load-bearing premise

The closed-form deterministic equivalent of the ergodic sum capacity remains accurate enough after clustering decisions are made that the relaxed solutions deliver the claimed fairness and rate performance on the true system.

What would settle it

On small networks where every possible clustering can be enumerated, compare the cluster assignment and true ergodic performance obtained from the continuous relaxation against the exact integer optimum to check whether they coincide.

Figures

Figures reproduced from arXiv: 2605.16960 by Boxiang Ren, Chaowen Deng, Hao Wu, Jie Fan, Jingchen Peng, Junyuan Wang, Ziyuan Lyu.

Figure 1
Figure 1. Figure 1: In Case 1, Jain’s index FI, minimum subnetwork sum ergodic capacity Cmin and total subnetwork sum ergodic capacity Csum with α = 0, 4, 8 and L = 150, M = 10. 100 125 150 175 200 K 0.86 0.88 0.90 0.92 0.94 FI α = 3 α = 7 α = 11 (a) 100 125 150 175 200 K 50 60 70 80 90 Cmin(bit/s/Hz) α = 3 α = 7 α = 11 (b) 100 125 150 175 200 K 1100 1200 1300 1400 1500 Csum(bit/s/Hz) α = 3 α = 7 α = 11 (c) [PITH_FULL_IMAGE:… view at source ↗
Figure 2
Figure 2. Figure 2: In Case 2, Jain’s index FI, minimum subnetwork sum ergodic capacity Cmin and total subnetwork sum ergodic capacity Csum with α = 3, 7, 11 and L = 150, M = 10. is employed to characterize the system’s minimum achievable performance, thereby reflecting the worst-case subnetwork state. A. Simulation Results Across Different Cases For different cases with varying values of α, we propose corresponding approache… view at source ↗
Figure 3
Figure 3. Figure 3: In Case 3, Jain’s index FI, minimum subnetwork sum ergodic capacity Cmin and total subnetwork sum ergodic capacity Csum with α = 0.5, 4.5, 8.5 and L = 150, M = 10. 100 125 150 175 200 K 0.86 0.88 0.90 0.92 0.94 0.96 FI α → ∞, AGP α → ∞, CGN (a) 100 125 150 175 200 K 50 60 70 80 90 Cmin(bit/s/Hz) α → ∞, AGP α → ∞, CGN (b) 100 125 150 175 200 K 1100 1200 1300 1400 Csum(bit/s/Hz) α → ∞, AGP α → ∞, CGN (c) [P… view at source ↗
Figure 4
Figure 4. Figure 4: In Case 4, Jain’s index FI, minimum subnetwork sum ergodic capacity Cmin and total subnetwork sum ergodic capacity Csum with proposed AGP and CGN [20] algorithms. L = 150, M = 10. respectively, while the overall quality decreases by only 2%. This highlights that a substantial improvement in fairness can be achieved with only a modest degradation in overall system performance. As for Case 2 and Case 3, we a… view at source ↗
Figure 5
Figure 5. Figure 5: Jain’s index FI, total subnetwork sum ergodic capacity [PITH_FULL_IMAGE:figures/full_fig_p012_5.png] view at source ↗
Figure 7
Figure 7. Figure 7: Per-subnetwork sum capacity comparison of clustered cell-free [PITH_FULL_IMAGE:figures/full_fig_p012_7.png] view at source ↗
read the original abstract

Clustered cell-free networking has emerged as a promising architecture to balance the high performance of cell-free massive MIMO and the scalability of traditional cellular systems. However, achieving fairness across subnetworks remains a critical yet largely unsolved challenge. This paper investigates the fairness problem in clustered cell-free networking and proposes a unified and tunable alpha-fairness scheme that effectively balances overall spectral efficiency and inter-subnetwork fairness. Using the closed-form deterministic equivalent of the ergodic sum capacity, we reformulate the combinatorial clustering problem as a continuous optimization problem. Leveraging the concavity/convexity properties of the alpha-fair objective, we classify the problem into four distinct cases according to the value of alpha. For each case, we establish the exact equivalence between the original integer program and its continuous relaxation, and develop efficient algorithms with guaranteed convergence. Extensive simulations show that the proposed scheme achieves up to 11% improvement in Jain's fairness index and 45% gain in minimum subnetwork capacity, with only a negligible 5% reduction in aggregate throughput.

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 addresses fairness in clustered cell-free massive MIMO by proposing a tunable α-fairness framework. It replaces the ergodic sum-capacity objective with its closed-form deterministic equivalent, converts the integer clustering problem into a continuous relaxation, and partitions the problem into four cases according to the value of α. For each case the authors claim to prove exact equivalence between the original integer program and the continuous relaxation, then supply convergent algorithms. Simulations report up to 11 % improvement in Jain’s fairness index and 45 % gain in minimum subnetwork rate at a 5 % cost in aggregate throughput.

Significance. If the deterministic-equivalent equivalence and convergence results hold and the approximation error does not materially distort the α-fair optimum, the work supplies a computationally tractable method for controlling the efficiency–fairness trade-off in scalable cell-free deployments. The explicit case-by-case analysis and guaranteed-convergence algorithms constitute concrete algorithmic contributions.

major comments (2)
  1. [§4] §4 (Equivalence proofs for the four α cases): the claimed exact equivalence is established between the integer program and its continuous relaxation when the objective is the deterministic equivalent of ergodic sum capacity. No error bound, uniform convergence result, or sensitivity analysis is supplied showing that a solution optimal for the approximated objective remains optimal (or near-optimal) for the true ergodic rates once clustering is fixed; in finite-dimensional regimes the approximation error can be non-uniform across subnetworks and directly affects the α-fairness metric.
  2. [§5] §5 (Simulation setup and evaluation): the optimization is performed with the deterministic equivalent while the reported fairness and rate gains are presumably evaluated with the true ergodic quantities. Without an explicit comparison of the two objectives on the same clustering solutions, it is impossible to verify that the claimed 11 % fairness and 45 % minimum-rate improvements survive the approximation gap.
minor comments (2)
  1. [§3] Notation for the deterministic equivalent (e.g., the definition of the effective channel matrix after clustering) is introduced without a self-contained reference to the underlying random-matrix result; a short appendix recalling the key steps would improve readability.
  2. [Fig. 3] Figure 3 caption does not state whether the plotted curves are obtained from the deterministic equivalent or from Monte-Carlo ergodic-rate evaluation.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments. We address each major comment below and indicate planned revisions to strengthen the manuscript.

read point-by-point responses

  1. Referee: [§4] §4 (Equivalence proofs for the four α cases): the claimed exact equivalence is established between the integer program and its continuous relaxation when the objective is the deterministic equivalent of ergodic sum capacity. No error bound, uniform convergence result, or sensitivity analysis is supplied showing that a solution optimal for the approximated objective remains optimal (or near-optimal) for the true ergodic rates once clustering is fixed; in finite-dimensional regimes the approximation error can be non-uniform across subnetworks and directly affects the α-fairness metric.

    Authors: We thank the referee for this observation. The exact equivalence we establish in Section 4 is between the integer clustering program and its continuous relaxation when the objective is the deterministic equivalent of the ergodic sum capacity; this is stated explicitly for each of the four α cases. The deterministic equivalent becomes exact in the large-system limit, but we acknowledge that no formal error bounds or sensitivity analysis for the true ergodic rates in finite dimensions are provided. Our simulations evaluate the final performance metrics with Monte Carlo estimates of the true ergodic rates, and the reported gains are observed under that evaluation. In the revised version we will add a short discussion of the approximation quality in Section 4 together with numerical comparisons of the deterministic-equivalent objective versus the true ergodic objective on the obtained clusterings. revision: partial

  2. Referee: [§5] §5 (Simulation setup and evaluation): the optimization is performed with the deterministic equivalent while the reported fairness and rate gains are presumably evaluated with the true ergodic quantities. Without an explicit comparison of the two objectives on the same clustering solutions, it is impossible to verify that the claimed 11 % fairness and 45 % minimum-rate improvements survive the approximation gap.

    Authors: We agree that an explicit side-by-side comparison would be helpful. The optimization in Section 5 uses the deterministic equivalent, while the fairness index and minimum-rate figures are computed from Monte Carlo estimates of the true ergodic rates. We will revise the manuscript to include a direct comparison (e.g., a table or additional plot) of the deterministic-equivalent objective value against the true ergodic sum capacity for the same clustering solutions, thereby quantifying the approximation gap and confirming that the reported improvements persist under the true objective. revision: yes

Circularity Check

0 steps flagged

No significant circularity; equivalence holds for reformulated deterministic-equivalent problem

full rationale

The derivation begins by invoking the closed-form deterministic equivalent of ergodic sum capacity as an external modeling tool to convert the combinatorial clustering task into a continuous optimization problem. For each of the four alpha cases the paper then proves exact equivalence between the resulting integer program and its continuous relaxation, followed by convergent algorithms. This equivalence is a mathematical property internal to the approximated objective and does not reduce to a self-definition, a fitted parameter renamed as prediction, or a load-bearing self-citation chain. The deterministic equivalent is treated as an independent input whose accuracy is an external modeling assumption rather than a derived claim; simulation results are presented separately as empirical validation. No quoted step collapses the claimed result to its own inputs by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the accuracy of the deterministic equivalent approximation and on the concavity/convexity properties that enable the four-case split; no new physical entities are introduced.

axioms (1)
  • domain assumption The closed-form deterministic equivalent accurately captures the ergodic sum capacity for the purpose of clustering optimization.
    Invoked to convert the combinatorial problem into a continuous one.

pith-pipeline@v0.9.0 · 5725 in / 1260 out tokens · 52970 ms · 2026-05-19T19:00:45.282466+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

51 extracted references · 51 canonical work pages

  1. [1]

    Towards load- balanced clustered cell-free networking: A tight relaxation approach,

    C. Deng, J. Fan, B. Ren, Z. Lyu, J. Wang, and H. Wu, “Towards load- balanced clustered cell-free networking: A tight relaxation approach,” in Proc. IEEE WCNC, Mar. 2025, pp. 1–6

    work page 2025

  2. [2]

    A vision of 6G wireless systems: Applications, trends, technologies, and open research problems,

    W. Saad, M. Bennis, and M. Chen, “A vision of 6G wireless systems: Applications, trends, technologies, and open research problems,”IEEE Network, vol. 34, no. 3, pp. 134–142, May/Jun. 2020

    work page 2020

  3. [3]

    Cell edge performance of cellular mobile systems,

    X. You, D. Wang, P. Zhu, and B. Sheng, “Cell edge performance of cellular mobile systems,”IEEE J. Sel. Areas Commun., vol. 29, no. 6, pp. 1139–1150, Jun. 2011

    work page 2011

  4. [4]

    Tse and P

    D. Tse and P. Viswanath,Fundamentals of Wireless Communication. United States of America by Cambridge University Press, New York, 2005

    work page 2005

  5. [5]

    Shannon-theoretic approach to a Gaussian cellular multiple- access channel,

    A. Wyner, “Shannon-theoretic approach to a Gaussian cellular multiple- access channel,”IEEE Trans. Inf. Theory, vol. 40, no. 6, pp. 1713–1727, Nov. 1994

    work page 1994

  6. [6]

    Coordinated multipoint: Concepts, performance, and field trial results,

    R. Irmer, H. Droste, P. Marsch, M. Grieger, G. Fettweis, S. Brueck, H.-P. Mayer, L. Thiele, and V . Jungnickel, “Coordinated multipoint: Concepts, performance, and field trial results,”IEEE Commun. Mag., vol. 49, no. 2, pp. 102–111, Feb. 2011

    work page 2011

  7. [7]

    Cell-free massive MIMO versus small cells,

    H. Q. Ngo, A. Ashikhmin, H. Yang, E. G. Larsson, and T. L. Marzetta, “Cell-free massive MIMO versus small cells,”IEEE Trans. Wireless Commun., vol. 16, no. 3, pp. 1834–1850, Jan. 2017

    work page 2017

  8. [8]

    Making cell-free massive MIMO competitive with MMSE processing and centralized implementation,

    E. Bj ¨ornson and L. Sanguinetti, “Making cell-free massive MIMO competitive with MMSE processing and centralized implementation,” IEEE Trans. Wireless Commun., vol. 19, no. 1, pp. 77–90, Jan. 2020

    work page 2020

  9. [9]

    Ubiquitous cell-free massive MIMO communications,

    G. Interdonato, E. Bj ¨ornson, H. Quoc Ngo, P. Frenger, and E. G. Larsson, “Ubiquitous cell-free massive MIMO communications,”EURASIP J. Wireless Commun. Networking, vol. 2019, no. 1, pp. 1–13, Aug. 2019

    work page 2019

  10. [10]

    Scalable cell-free massive MIMO systems,

    E. Bj ¨ornson and L. Sanguinetti, “Scalable cell-free massive MIMO systems,”IEEE Trans. Commun., vol. 68, no. 7, pp. 4247–4261, Jul. 2020

    work page 2020

  11. [11]

    Cluster-wise processing in fronthaul-aware cell-free massive MIMO systems,

    Z. Mobini, A. H. Gokceoglu, L. Wang, G. Peters, H. Shin, and H. Q. Ngo, “Cluster-wise processing in fronthaul-aware cell-free massive MIMO systems,”IEEE Trans. Wireless Commun., pp. 1–1, Oct. 2025

    work page 2025

  12. [12]

    Optimal decomposition for large-scale infrastructure- based wireless networks,

    L. Dai and B. Bai, “Optimal decomposition for large-scale infrastructure- based wireless networks,”IEEE Trans. Wireless Commun., vol. 16, no. 8, pp. 4956–4969, Aug. 2017

    work page 2017

  13. [13]

    Clustered cell-free networking: A graph partitioning approach,

    J. Wang, L. Dai, L. Yang, and B. Bai, “Clustered cell-free networking: A graph partitioning approach,”IEEE Trans. Wireless Commun., vol. 22, no. 8, pp. 5349–5364, Aug. 2023

    work page 2023

  14. [14]

    Rate-constrained network decomposition for clustered cell-free networking,

    J. Wang, L. Dai, L. Yang, and B. Bai, “Rate-constrained network decomposition for clustered cell-free networking,” inProc. IEEE ICC, May 2022, pp. 2549–2554

    work page 2022

  15. [15]

    A sequential minK-cut approach for sum rate maximization of clustered cell-free networking,

    B. Ren, C. Deng, H. Hao, H. Wu, and J. Wang, “A sequential minK-cut approach for sum rate maximization of clustered cell-free networking,” inProc. IEEE ICC, Jun. 2024, pp. 4900–4905

    work page 2024

  16. [16]

    Tunable weighted kernelk-means for clustered cell-free networking acceleration and beam on-off control,

    X. Zeng, J. Wang, K. Yue, M. Dong, and B. Bai, “Tunable weighted kernelk-means for clustered cell-free networking acceleration and beam on-off control,” inProc. IEEE ICC, June 2024, pp. 4311–4316

    work page 2024

  17. [17]

    Tight differ- entiable relaxation of sum ergodic capacity maximization for clustered cell-free networking,

    B. Ren, H. Hao, Z. Lyu, J. Peng, J. Wang, and H. Wu, “Tight differ- entiable relaxation of sum ergodic capacity maximization for clustered cell-free networking,” inProc. IEEE ISIT, Jul. 2024, pp. 2448–2453

    work page 2024

  18. [18]

    Optimizing clustered cell-free networking for sum ergodic capacity maximization with joint processing constraint,

    F. Xia, J. Wang, and L. Dai, “Optimizing clustered cell-free networking for sum ergodic capacity maximization with joint processing constraint,” IEEE Trans. Wireless Commun., vol. 24, no. 1, pp. 571–584, Jan. 2025

    work page 2025

  19. [19]

    C 2: A capacity-centric architecture toward future wireless networking,

    L. Yang, P. Li, M. Dong, B. Bai, D. Zaporozhets, X. Chen, W. Han, and B. Li, “C 2: A capacity-centric architecture toward future wireless networking,”IEEE Trans. Wireless Commun., vol. 21, no. 10, pp. 8134– 8147, Oct. 2022

    work page 2022

  20. [20]

    CGN: A capacity-guaranteed network architecture for future ultra-dense wireless systems,

    C. Deng, L. Yang, H. Wu, D. Zaporozhets, M. Dong, and B. Bai, “CGN: A capacity-guaranteed network architecture for future ultra-dense wireless systems,” inProc. IEEE ICC, May 2022, pp. 1853–1858

    work page 2022

  21. [21]

    Fair end-to-end window-based congestion control,

    J. Mo and J. Walrand, “Fair end-to-end window-based congestion control,”IEEE/ACM Trans. Networking, vol. 8, no. 5, pp. 556–567, Oct. 2000

    work page 2000

  22. [22]

    A quantitative measure of fairness and discrimination,

    R. K. Jain, D.-M. W. Chiu, W. R. Haweet al., “A quantitative measure of fairness and discrimination,”Eastern Research Laboratory, Digital Equipment Corporation, Hudson, MA, vol. 21, no. 1, pp. 2022–2023, 1984

    work page 2022

  23. [23]

    Deterministic equivalents for certain functionals of large random matrices,

    W. Hachem, P. Loubaton, and J. Najim, “Deterministic equivalents for certain functionals of large random matrices,”Ann. Appl. Probab., vol. 17, no. 3, p. 875–930, June 2007

    work page 2007

  24. [24]

    Optimal power allocation scheme for non- orthogonal multiple access withα-fairness,

    P. Xu and K. Cumanan, “Optimal power allocation scheme for non- orthogonal multiple access withα-fairness,”IEEE J. Sel. Areas Com- mun., vol. 35, no. 10, pp. 2357–2369, Oct. 2017

    work page 2017

  25. [25]

    Enhancing user fairness in wireless powered communication networks with STAR-RIS,

    G. Zhu, X. Mu, L. Guo, A. Huang, and S. Xu, “Enhancing user fairness in wireless powered communication networks with STAR-RIS,”IEEE Internet Things J., vol. 12, no. 3, pp. 2659–2673, Feb. 2025

    work page 2025

  26. [26]

    Fairness scheduling in user-centric cell-free massive MIMO wireless networks,

    F. G ¨ottsch, N. Osawa, I. Kanno, T. Ohseki, and G. Caire, “Fairness scheduling in user-centric cell-free massive MIMO wireless networks,” IEEE Trans. Wireless Commun., vol. 23, no. 9, pp. 11 942–11 957, Sep. 2024

    work page 2024

  27. [27]

    Fairness in wire- less networks:issues, measures and challenges,

    H. SHI, R. V . Prasad, E. Onur, and I. Niemegeers, “Fairness in wire- less networks:issues, measures and challenges,”IEEE Commun. Surv. Tutorials, vol. 16, no. 1, pp. 5–24, First Quarter 2014

    work page 2014

  28. [28]

    Fairness for non-orthogonal multiple access in 5G systems,

    S. Timotheou and I. Krikidis, “Fairness for non-orthogonal multiple access in 5G systems,”IEEE Signal Process Lett., vol. 22, no. 10, pp. 1647–1651, Oct. 2015

    work page 2015

  29. [29]

    On the outage performance of non-orthogonal multiple access with 1-bit feedback,

    P. Xu, Y . Yuan, Z. Ding, X. Dai, and R. Schober, “On the outage performance of non-orthogonal multiple access with 1-bit feedback,” IEEE Trans. Wireless Commun., vol. 15, no. 10, pp. 6716–6730, Oct. 2016

    work page 2016

  30. [30]

    Fairness of user clustering in MIMO non-orthogonal multiple access systems,

    Y . Liu, M. Elkashlan, Z. Ding, and G. K. Karagiannidis, “Fairness of user clustering in MIMO non-orthogonal multiple access systems,”IEEE Commun. Lett., vol. 20, no. 7, pp. 1465–1468, Jul. 2016

    work page 2016

  31. [31]

    Proportional fairness-aware task scheduling in space-air-ground integrated networks,

    G. Sun, Y . Wang, H. Yu, and M. Guizani, “Proportional fairness-aware task scheduling in space-air-ground integrated networks,”IEEE Trans. Serv. Comput., vol. 17, no. 6, pp. 4125–4137, Nov.-Dec. 2024. 15

    work page 2024

  32. [32]

    An axiomatic theory of fairness in network resource allocation,

    T. Lan, D. Kao, M. Chiang, and A. Sabharwal, “An axiomatic theory of fairness in network resource allocation,” inProc. IEEE INFOCOM, Mar. 2010, pp. 1–9

    work page 2010

  33. [33]

    Rate control for com- munication networks: shadow prices, proportional fairness and stability,

    F. P. Kelly, A. K. Maulloo, and D. K. H. Tan, “Rate control for com- munication networks: shadow prices, proportional fairness and stability,” Journal of the Operational Research society, vol. 49, no. 3, pp. 237–252, Mar. 1998

    work page 1998

  34. [34]

    Foundations of user- centric cell-free massive MIMO,

    ¨O. T. Demir, E. Bj ¨ornson, L. Sanguinettiet al., “Foundations of user- centric cell-free massive MIMO,”Found. Trends Signal Process., vol. 14, no. 3-4, pp. 162–472, 2021

    work page 2021

  35. [35]

    Graph matching by simplified convex-concave relaxation procedure,

    Z. Liu, H. Qiao, X. Yang, and S. C. Hoi, “Graph matching by simplified convex-concave relaxation procedure,”Int. J. Comput. Vision, vol. 109, pp. 169–186, 2014

    work page 2014

  36. [36]

    GNCCP—graduated nonconvexityand concavity procedure,

    Z. Liu and H. Qiao, “GNCCP—graduated nonconvexityand concavity procedure,”IEEE Trans. Pattern Anal. Mach. Intell., vol. 36, no. 6, pp. 1258–1267, 2013

    work page 2013

  37. [37]

    An algorithm for quadratic programming,

    M. Frank, P. Wolfeet al., “An algorithm for quadratic programming,” Naval research logistics quarterly, vol. 3, no. 1-2, pp. 95–110, 1956

    work page 1956

  38. [38]

    The optimal production transport: Model and algorithm,

    F. Jie, T. Wu, and H. Wu, “The optimal production transport: Model and algorithm,”East Asian Journal on Applied Mathematics, vol. 15, no. 4, pp. 787–816, 2025

    work page 2025

  39. [39]

    An efficient algorithm for nonconvex-linear minimax optimization problem and its application in solving weighted maximin dispersion problem,

    W. Pan, J. Shen, and Z. Xu, “An efficient algorithm for nonconvex-linear minimax optimization problem and its application in solving weighted maximin dispersion problem,”Comput. Optim. Appl., vol. 78, pp. 287– 306, 2021

    work page 2021

  40. [40]

    S. P. Boyd and L. Vandenberghe,Convex optimization. Cambridge university press, 2004

    work page 2004

  41. [41]

    Linear assignment problems,

    S. Martello and P. Toth, “Linear assignment problems,” inNorth-Holland Mathematics Studies. Elsevier, 1987, vol. 132, pp. 259–282

    work page 1987

  42. [42]

    The frank-wolfe algorithm: a short introduction,

    S. Pokutta, “The frank-wolfe algorithm: a short introduction,”Jahres- bericht der Deutschen Mathematiker-Vereinigung, vol. 126, no. 1, pp. 3–35, 2024

    work page 2024

  43. [43]

    Nonlinear programming,

    H. W. Kuhn and A. W. Tucker, “Nonlinear programming,” inTraces and emergence of nonlinear programming. Springer, 2013, pp. 247–258

    work page 2013

  44. [44]

    Sinkhorn distances: Lightspeed computation of optimal transport,

    M. Cuturi, “Sinkhorn distances: Lightspeed computation of optimal transport,”Advances in neural information processing systems, vol. 26, 2013

    work page 2013

  45. [45]

    Balanced k-means for clustering,

    M. I. Malinen and P. Fr ¨anti, “Balanced k-means for clustering,” in Structural, Syntactic, and Statistical Pattern Recognition: Joint IAPR International Workshop, S+ SSPR 2014, Joensuu, Finland, August 20- 22, 2014. Proceedings. Springer, 2014, pp. 32–41

    work page 2014

  46. [46]

    Solving a class of non-convex min-max games using iterative first order methods,

    M. Nouiehed, M. Sanjabi, T. Huang, J. D. Lee, and M. Razaviyayn, “Solving a class of non-convex min-max games using iterative first order methods,” inProc. NeurIPS, vol. 32, Dec. 2019

    work page 2019

  47. [47]

    The theory of max-min, with applications,

    J. M. Danskin, “The theory of max-min, with applications,”SIAM J. Appl. Math, vol. 14, no. 4, pp. 641–664, 1966

    work page 1966

  48. [48]

    Nonconvex min-max optimization: Applications, challenges, and recent theoretical advances,

    M. Razaviyayn, T. Huang, S. Lu, M. Nouiehed, M. Sanjabi, and M. Hong, “Nonconvex min-max optimization: Applications, challenges, and recent theoretical advances,”IEEE Signal Process Mag., vol. 37, no. 5, pp. 55–66, Sep. 2020

    work page 2020

  49. [49]

    A method for finding projections onto the intersection of convex sets in Hilbert spaces,

    J. P. Boyle and R. L. Dykstra, “A method for finding projections onto the intersection of convex sets in Hilbert spaces,” inAdvances in Order Restricted Statistical Inference. Springer, 1986, pp. 28–47

    work page 1986

  50. [50]

    Dykstra’s algorithm with Bregman projections: A convergence proof,

    H. H. Bauschke and A. S. Lewis, “Dykstra’s algorithm with Bregman projections: A convergence proof,”Optimization, vol. 48, no. 4, pp. 409– 427, 2000

    work page 2000

  51. [51]

    Two Bregman projection methods for solving variational inequalities,

    D. V . Hieu and S. Reich, “Two Bregman projection methods for solving variational inequalities,”Optimization, vol. 71, no. 7, pp. 1777–1802, 2022

    work page 2022