Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

SIG-SDP: Sparse Interference Graph-Aided Semidefinite Programming for Large-Scale Wireless Time-Sensitive Networking

T0 review · 3 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read The paper claims that sparse interference graphs can make semidefinite-programming-based slot assignment practical for large wireless time-sensitive networks, cutting computation to near-linear in the number of users while preserving or…

desk verdict A solid sparse SDP heuristic for WTSN scheduling, but the exact-minimum claim rests on an unproven rounding oracle, and the paper's own reliability numbers miss the stated target. read the letter →

arxiv 2501.11307 v2 pith:MX57JJSV submitted 2025-01-20 eess.SP

Zhouyou Gu , Jihong Park , Branka Vucetic , Jinho Choi This is my paper · ORCID
classification eess.SP MSC 90C2290C2705C15
keywords time-sensitivenetworkingsemidefiniteprogramminginterferencegraphslotschedulingmatrixmultiplicativeweightssparsitybinarysearchwirelessreliability
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

The paper aims to make semidefinite-programming based slot assignment practical in large wireless time-sensitive networks, where the goal is to minimize the number of time slots while keeping interference below reliability thresholds. Because the problem is NP-hard, the authors relax it to an SDP over a positive semidefinite matrix that encodes, for each pair of users, how likely they share a slot. Their SIG-SDP framework exploits the natural sparsity of interference graphs to bound the search range for the minimum slot count, solves each feasibility check with a matrix multiplicative weights solver that only touches interfering user pairs, and rounds the relaxed solution back to a slot assignment. If correct, this reduces the runtime from high-degree polynomial to near-linear in the number of users, about ten times faster than the classical ADMM solver, and lowers packet loss rates by up to two orders of magnitude relative to linear relaxation and heuristic baselines.

What carries the argument

The central object is the PSD matrix $X$, whose $(i,j)$-entry represents how likely users $i$ and $j$ are assigned the same slot: slot choices are mapped to unit radial vectors of a regular simplex so that the indicator $\mathbf{1}\{z_i = z_j\}$ becomes $\frac{1}{Z}(1 + (Z-1)X_{ij})$. The argument is carried by three pieces: sparse interference graphs $G_{\mathrm{asso}}$ and $G_{\mathrm{intp}}$ whose edge sets bound the number of constraints and give chromatic-number bounds on $Z^*$; the matrix multiplicative weights algorithm, which plays a two-player game where dual weights follow the hedge rule and primal iterates come from a sketched matrix exponential; and a random-vector rounding step that converts the relaxed PSD matrix into an integer slot assignment. Sparsity is what turns an otherwise high-polynomial SDP into near-linear work, because only non-zero constraint entries are updated.

What would settle it

Take a small network instance where the true minimum slot count is known by exhaustive search, run Algorithm 1, and check whether the binary search ever raises the lower bound because the rounding fails to assign all users even though a feasible $Z$-slot assignment exists. If such an instance is found, the framework returns a slot count above the true minimum.

Watch

Extended reading notes

Core claim

On its own terms, the paper establishes that the NP-hard slot-minimization problem can be approached by a binary search whose bounds come from graph coloring: Theorem 1 gives $Z^*$ between $1 - \lambda_{\max}(Q)/\lambda_{\min}(Q)$ and $\chi(G_{\mathrm{intp}}) \le \Omega+1$, using the association graph and the interference-power graph. For a fixed candidate $Z$, the integer feasibility problem is relaxed into an SDP constraint satisfaction problem over positive semidefinite matrices with unit diagonal, and the paper proves strong duality for this formulation and a convergence bound for the matrix multiplicative weights solver: with $N = \eta^{-2}(\ln K + \ln C)$ iterations the duality gap is $O(\eta K)$. The implementation uses randomized sketching to approximate the matrix exponential, keeps only elements of the PSD matrix corresponding to graph edges, and reports a near-linear overall complexity $O(N K \Omega^2 \log \Omega)$. Simulations show convergence in roughly 100 iterations across network sizes and a slot-count minimum that matches or beats the ADMM-based framework while running up to ten times faster.

Load-bearing premise

The load-bearing premise is that when a feasible slot assignment with $Z$ slots exists, the relaxed SDP solution plus the random rounding step will actually find one; the paper does not prove this, so the returned minimum slot count could be larger than the true minimum.

Editorial extensions

If this is right

  • The minimum number of slots can be located by binary search inside $[1 - \lambda_{\max}(Q)/\lambda_{\min}(Q),\, \Omega+1]$, so the search range shrinks with graph sparsity rather than growing with the number of users.
  • The MMW solver certifies a duality gap of $O(\eta K)$ after $O(\eta^{-2}(\ln K + \ln C))$ iterations, giving a practical stopping rule for the relaxed feasibility problem.
  • Exploiting edge sparsity makes per-iteration cost $O(K\Omega^2)$, which is near-linear in $K$ whenever the maximum interference-neighbor count $\Omega$ saturates as the network grows.
  • On the simulated WTSN scenarios, the framework returns the minimum slot count with about 10 times less computation than the ADMM-based SDP and packet error rates roughly 100 times lower than linear relaxation and 10 times lower than greedy heuristics.
  • In the online architecture, updating assignments during user mobility beats the low-complexity heuristic for slow-moving users, and the advantage shrinks as mobility increases.

Reading between the lines

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

  • The paper does not prove that the random rounding in (25)-(26) certifies feasibility, so the returned $Z$ may overestimate the true minimum; a testable extension is to replace rounding with an exact feasibility check for small instances and compare the binary-search verdicts.
  • Because Theorem 3 assumes exact matrix-exponential sketching, the finite-rank approximation error enters as an unquantified constant factor; one testable design question is how $D$ must scale with $K$ or $\Omega$ to preserve the $O(\eta K)$ gap in practice.
  • The sparsity assumption itself is geometric: with finite receiver sensitivity $\gamma$, each user's interfering neighborhood is bounded in area, so $\Omega$ saturates. In fading channels the graph edges become stochastic, so an extension would be to build a random graph model and ask whether expected chromatic-number bounds still hold.
  • The online results suggest a complexity-reliability tradeoff: further acceleration, such as one-shot graph-to-assignment mapping, would directly improve reliability under mobility, a direction the paper names in its conclusion.
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.

Referee Report

3 major / 5 minor

Summary. The paper studies the NP-hard problem of minimizing the number of time slots in a wireless time-sensitive network under association and interference constraints. It proposes SIG-SDP, a framework that relaxes the integer slot-assignment problem to an SDP constraint satisfaction problem, bounds the optimal slot count using chromatic numbers of sparse interference graphs, and performs a binary search in which each candidate slot count is tested by solving the relaxed SDP with a matrix multiplicative weights (MMW) algorithm and then rounding the solution to integer assignments. The authors also present an online architecture for dynamic networks and report simulations showing near-linear scaling, faster computation than ADMM, and lower packet error rates than heuristics and linear relaxation.

Significance. If the claims hold, the paper would make SDP-based interference management practical for networks with hundreds of users, a useful step for WTSN. The paper has clear strengths: a well-specified system model, a thoughtful exploitation of interference graph sparsity, a complete MMW formulation with a duality-gap convergence theorem, extensive simulations across network sizes, and a promised open-source implementation. The main significance hinges on the exact-slot-minimization claim, which currently rests on an unproven randomized-rounding feasibility oracle, and on the reliability claim, which is directly qualified by the paper's own simulation results.

major comments (3)
  1. [Section IV-B, Algorithm 1 (lines 8-9)] The binary search treats the randomized rounding in (25)-(26) as a feasibility oracle: if rounding succeeds for a tested Z, the upper bound is lowered; if it fails, the lower bound is raised to Z+1. This is valid only if rounding failure implies that no integer assignment with Z slots exists. The paper proves no such implication. The SDP relaxation (23) is a necessary condition for integer feasibility, but there is no theorem, analogous to the Goemans-Williamson rounding guarantee, showing that whenever an integer assignment with Z slots exists, the random-vector rounding of an optimal or near-optimal SDP solution will find one. Consequently, the algorithm can return a slot count strictly larger than the true optimum, and the abstract's claim that the framework "minimizes the number of slots" is not established. This is structurally independent of Theorem 3, which only bounds the duality gap of the SDP solution. The authors should either prove a rounding-success guarantee (with high probability) or explicitly reframe the method as a heuristic that returns an upper bound on the minimum number of slots.
  2. [Section VII-E, Fig. 10a] The paper's own simulation results show that approximately 5-10% of users have packet error rates exceeding the required threshold of 10^-5, and the text attributes this to unmeasurable interference that is not included in the constraints. This directly weakens the abstract's claim of "ensuring reliability subject to interference constraints": the constraints in (10) are based only on measured interference, so the method does not guarantee the actual packet error rate requirement. The authors should qualify the reliability claim throughout the paper to refer to constraints on measured interference, and should prominently state that the end-to-end reliability target is not guaranteed when unmeasurable interference is present.
  3. [Section V-D, Theorem 3] Theorem 3 is proved under the explicit assumption that the randomized sketching in (39)-(42) is exact. The paper then states, without proof, that sketching error only scales the gap by a constant factor, citing [33]. This is a load-bearing step for the convergence claim in the practical algorithm, and the conditions under which the constant-factor argument applies to this specific SDP CSP (with the constraint structure and the fixed sketch dimension D) should be spelled out. Additionally, the near-linear complexity claim O(K) depends on treating both the MMW iteration count N and the maximum degree Omega as constants; Theorem 3 gives N = O((ln K + ln C)/eta^2), which grows logarithmically, and the constancy of Omega is an empirical observation, not a proven bound. The paper should clearly separate the proven asymptotic statement from the simulation-supported constant-N and constant-Omega assumptions.
minor comments (5)
  1. [Equation (39)-(42)] The notation for the sketching approximation is confusing: in (39), v_k are rows of the sketched product, but in (40)-(42) they are used both as row vectors and as factors in an outer product; the sum in (41) should be over the norms of all K rows. Please define the dimensions of V[n] and the resulting matrix consistently.
  2. [Algorithm 1] The for-loop "for m = 1, . . . , do" has no explicit termination condition except the break at line 10; this is not a formal loop structure. Please specify a while-loop condition such as "while Z_a < Z_b".
  3. [Figure 9] The axis labeling in Fig. 9 is unclear: the y-axis appears to list values of eta, but the caption says the x-axis is the number of iterations. Please clarify what the color/height represents and label both axes explicitly.
  4. [Reference [33]] The author list of [33] is garbled: "Y. Carmon, J. C. Duchi, S. Aaron, and T. Kevin" should be "Y. Carmon, J. C. Duchi, A. Sidford, and K. Tian."
  5. [Section IV-B, Eq. (24)-(26)] In (45), the returned value U(Σ)^{1/2} is a K by D matrix, but it is called a gram form of a K by K matrix. Please clarify the dimensions and how the rounding in (25)-(26) uses this rectangular factor.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: the bounds, SDP relaxation, and MMW convergence rely on external graph-coloring, minmax, and regret-bound results; the rounding-oracle gap is a correctness risk, not circularity.

full rationale

The paper's chain of derivation is self-contained against standard external results rather than reducing to its own inputs. Theorem 1 bounds the optimal slot number using the Hoffman eigenvalue bound and greedy coloring of graph theory, citing standard texts. The SDP relaxation in (23) is a standard continuous relaxation of the integer CSP (15), with the PSD matrix representing pairwise same-slot likelihood; this is a modeling choice, not a definition that forces the claimed minimum. The MMW solver's convergence (Theorem 3) is proved by combining the Arora-Kale matrix multiplicative weights regret bound and the Freund-Schapire hedge rule bound, both external and machine-checkable in the literature. The claimed slot-number minimization does rest on Algorithm 1 using randomized rounding (25)-(26) as a feasibility oracle, where rounding failure triggers Z_a = Z+1; but no theorem proves rounding failure implies integer infeasibility. This is a genuine correctness gap in the exact-minimization claim, but it is not circularity: the rounding outcome is not definitionally equivalent to feasibility, and the paper does not derive feasibility from its own output. The hyperparameters eta, N, and D are tuned on the same simulations that demonstrate convergence, and N is fixed as a constant in the complexity claim based on those simulations; this is an empirical self-supporting loop and a reproducibility concern, but it is not a fitted parameter renamed as a prediction, since the performance comparisons against ADMM, LP, and heuristics are not statistically forced by that tuning. Self-citations to [17] and [24] appear only as background for SDP-based assignment and scheduling and are not load-bearing; the core solver and bounds rely on external citations [31]-[33], [36], [37]. Therefore no circular step can be exhibited under the required standard.

Assumptions & free parameters 3 free parameters · 7 assumptions · 0 invented entities

The central method rests on standard SDP and MMW machinery plus several domain assumptions. The load-bearing free parameters are algorithm hyperparameters tuned to the simulations. The most important ad hoc premise is that randomized rounding can serve as a feasibility oracle for the binary search, which is assumed without proof. No invented physical entities are introduced.

free parameters (3)
  • MMW step size η = 0.04
    Chosen from convergence sweeps in Section VII-D; not derived from Theorem 3, which for the simulated K and C would require a much smaller η or larger N to make the O(ηK) bound small.
  • MMW iteration count N = 150
    Set empirically after Fig. 9; Theorem 3 prescribes N = (1/η²)(ln K + ln C), which is far larger for K=675, C≈6000. The paper relies on observed convergence, not the theorem.
  • Sketch dimension D = D = ⌈2(Z-1)⌉
    The paper sets β=2 in Section VII-A and argues D=O(Z) suffices; no rigorous error bound is given for this choice.
assumptions (7)
  • standard math A slot choice can be represented as a unit radial vector of a regular simplex, with inner products given by (16).
    Used in Section IV-A to derive the indicator function (18) and the SDP relaxation; this is standard geometry.
  • domain assumption Interference below the receiver sensitivity threshold γ is set to zero in the constraints (7).
    Section II-B; the actual packet errors in simulation include this unmeasurable interference, so the constraints may not guarantee the stated 1e-5 reliability.
  • domain assumption Each BS decodes only one associated user per slot.
    Section II-A; this produces the hard constraint (9).
  • domain assumption The channel is deterministic and without fading, so path gains can be measured exactly.
    Stated in the conclusion (Section VIII); the stochastic graph case is deferred.
  • ad hoc to paper Randomized rounding of the relaxed PSD solution is a correct feasibility oracle for the binary search.
    Algorithm 1 lines 7-9; no theorem links SDP feasibility to rounding success.
  • ad hoc to paper The neighborhood bound Ω and MMW iteration count N are effectively constant as K grows.
    Used to claim near-linear complexity in Section V-E; supported only by Fig. 7 and an informal geometric argument, not a theorem.
  • ad hoc to paper Randomized sketching with D=O(Z) approximates the matrix exponential well enough for the convergence claim.
    Section V-C-D; Theorem 3 assumes the exponential is exact, while the implementation uses (39)-(42).

how reviews work

0 comments
Cite this review

Pith. "Pith review of SIG-SDP: Sparse Interference Graph-Aided Semidefinite Programming for Large-Scale Wireless Time-Sensitive Networking." pith.science (2026). https://pith.science/paper/MX57JJSV

@misc{pith2026250111307,
  author       = {Pith},
  title        = {Pith review of: SIG-SDP: Sparse Interference Graph-Aided Semidefinite Programming for Large-Scale Wireless Time-Sensitive Networking},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MX57JJSV}},
  note         = {Machine review of arXiv:2501.11307}
}
read the original abstract

Wireless time-sensitive networking (WTSN) is essential for Industrial Internet of Things. We address the problem of minimizing time slots needed for WTSN transmissions while ensuring reliability subject to interference constraints -- an NP-hard task. Existing semidefinite programming (SDP) methods can relax and solve the problem but suffer from high polynomial complexity. We propose a sparse interference graph-aided SDP (SIG-SDP) framework that exploits the interference's sparsity arising from attenuated signals between distant user pairs. First, the framework utilizes the sparsity to establish the upper and lower bounds of the minimum number of slots and uses binary search to locate the minimum within the bounds. Here, for each searched slot number, the framework optimizes a positive semidefinite (PSD) matrix indicating how likely user pairs share the same slot, and the constraint feasibility with the optimized PSD matrix further refines the slot search range. Second, the framework designs a matrix multiplicative weights (MMW) algorithm that accelerates the optimization, achieved by only sparsely adjusting interfering user pairs' elements in the PSD matrix while skipping the non-interfering pairs. We also design an online architecture to deploy the framework to adjust slot assignments based on real-time interference measurements. Simulations show that the SIG-SDP framework converges in near-linear complexity and is highly scalable to large networks. The framework minimizes the number of slots with up to 10 times faster computation and up to 100 times lower packet loss rates than compared methods. The online architecture demonstrates how the algorithm complexity impacts dynamic networks' performance.

Figures

Figures reproduced from arXiv: 2501.11307 by the authors.

Figure 1
Figure 1. Illustration of a WTSN system for IIoT applications. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Illustration of a WTSN system. two matrices are denoted as X • Y = P i P j Xi,jYi,j . We denote the maximum and minimum eigenvalues of a matrix X as λmax(X) and λmin(X), respectively. The ℓ1 and ℓ2 norm of a vector x is denoted as |x| and ∥x∥, respectively. The spectral norm of a matrix X is denoted as ∥X∥, which equals the maximum absolute value of eigenvalues of X when X is symmetric. I K is the K × K identity mat… view at source ↗
Figure 3
Figure 3. A graphical illustration of WTSN: a) an interference network topology [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: Unit radial vectors in regular (Z−1)-simplexes, with Z = 2, 3, 4, . . . , representing all possible slot choices for given Z. IV. RELAXATION AND SPARSITY-AWARE BINARY SEARCH IN SIG-SDP This section first presents the SDP relaxation of the slot as￾signment problem in (1…
Figure 5
Figure 5. Figure 5: Illustration of the matrix multiplicative weights algorithm as a turn [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Interaction between the solver and the network in time. [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 8
Figure 8. Figure 8: Sparsity patterns of non-zero elements in constraint coefficient [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]
Figure 9
Figure 9. Figure 9: The convergence of the duality gap in the MMW, [PITH_FULL_IMAGE:figures/full_fig_p012_9.png]
Figure 10
Figure 10. Figure 10: CDFs of the minimum number of slots and the error rates when [PITH_FULL_IMAGE:figures/full_fig_p012_10.png]
Figure 11
Figure 11. Figure 11: User packet error rates when using different schemes. [PITH_FULL_IMAGE:figures/full_fig_p013_11.png]
Figure 13
Figure 13. Figure 13: Average packet error rates when using the proposed SIG-SDP [PITH_FULL_IMAGE:figures/full_fig_p014_13.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Scalable Interference Graph Learning for Low-Latency Wi-Fi Networks using Hashing-based Evolution Strategy

    eess.SP 2025-02 conditional novelty 5.0 of 10

    A neural graph model trained with evolution strategies and deep hashing assigns Wi-Fi 7 scheduled transmission slots, using about 25% fewer slots and losing up to 30% fewer packets in 1,000-device simulations.

Reference graph

Works this paper leans on

49 extracted references · 40 canonical work pages · cited by 1 Pith paper

  1. [33]

    A rank-1 sketch for matrix multiplicative weights,

    Y . Carmon, J. C. Duchi, S. Aaron, and T. Kevin, “A rank-1 sketch for matrix multiplicative weights,” in Conference on Learning Theory . PMLR, 2019, pp. 589–623

    work page 2019

  2. [1]

    Indus- trial Internet of Things: Challenges, opportunities, and directions,

    E. Sisinni, A. Saifullah, S. Han, U. Jennehag, and M. Gidlund, “Indus- trial Internet of Things: Challenges, opportunities, and directions,” IEEE Transactions on Industrial Informatics , vol. 14, no. 11, pp. 4724–4734, 2018

    work page 2018

  3. [2]

    Industrial communication systems and their future challenges: Next-generation Ethernet, IIoT, and 5G,

    S. Vitturi, C. Zunino, and T. Sauter, “Industrial communication systems and their future challenges: Next-generation Ethernet, IIoT, and 5G,” Proceedings of the IEEE , vol. 107, no. 6, pp. 944–961, 2019

    work page 2019

  4. [3]

    A comprehensive survey on interoperability for IIoT: Taxonomy, standards, and future directions,

    A. Hazra, M. Adhikari, T. Amgoth, and S. N. Srirama, “A comprehensive survey on interoperability for IIoT: Taxonomy, standards, and future directions,” ACM Computing Surveys (CSUR) , vol. 55, no. 1, pp. 1–35, 2021

    work page 2021

  5. [4]

    Ultra-low latency (ULL) networks: The IEEE TSN and IETF DetNet standards and related 5G ULL research,

    A. Nasrallah, A. S. Thyagaturu, Z. Alharbi, C. Wang, X. Shao, M. Reisslein, and H. ElBakoury, “Ultra-low latency (ULL) networks: The IEEE TSN and IETF DetNet standards and related 5G ULL research,” IEEE Communications Surveys & Tutorials , vol. 21, no. 1, pp. 88–145, 2018

    work page 2018

  6. [5]

    Service requirements for cyber-physical control applications in vertical domains,

    3GPP, “Service requirements for cyber-physical control applications in vertical domains,” 3GPP, TS 22.104, 2018, v16.0.0

    work page 2018

  7. [6]

    A comprehensive survey of wireless time-sensitive networking (tsn): Architecture, technologies, applications, and open issues,

    K. Zanbouri, M. Noor-A-Rahim, J. John, C. J. Sreenan, H. V . Poor, and D. Pesch, “A comprehensive survey of wireless time-sensitive networking (tsn): Architecture, technologies, applications, and open issues,” IEEE Communications Surveys & Tutorials , 2024

    work page 2024

  8. [7]

    5G industrial networks with CoMP for URLLC and time sensitive network architecture,

    M. Khoshnevisan, V . Joseph, P. Gupta, F. Meshkati, R. Prakash, and P. Tinnakornsrisuphap, “5G industrial networks with CoMP for URLLC and time sensitive network architecture,” IEEE Journal on Selected Areas in Communications , vol. 37, no. 4, pp. 947–959, 2019. 16

    work page 2019

Show all 49 references
  1. [8]

    w-SHARP: Implementation of a high-performance wireless time-sensitive network for low latency and ultra-low cycle time industrial applications,

    ´O. Seijo, J. A. L´opez-Fern´andez, and I. Val, “w-SHARP: Implementation of a high-performance wireless time-sensitive network for low latency and ultra-low cycle time industrial applications,” IEEE Transactions on Industrial Informatics, vol. 17, no. 5, pp. 3651–3662, 2020

    2020

  2. [9]

    Wireless time sensitive networking impact on an industrial collab- orative robotic workcell,

    S. Sudhakaran, K. Montgomery, M. Kashef, D. Cavalcanti, and R. Can- dell, “Wireless time sensitive networking impact on an industrial collab- orative robotic workcell,” IEEE Transactions on Industrial Informatics , vol. 18, no. 10, pp. 7351–7360, 2022

    2022

  3. [10]

    Scheduling for time-critical applications utilizing tcp in software- based 802.1 qbv wireless tsn,

    R. Candell, K. Montgomery, M. K. Hany, S. Sudhakaran, and D. Caval- canti, “Scheduling for time-critical applications utilizing tcp in software- based 802.1 qbv wireless tsn,” in 2023 IEEE 19th International Con- ference on Factory Communication Systems (WFCS) . IEEE, 2023, pp. 1–8

    2023

  4. [11]

    Knowledge-assisted deep reinforcement learning in 5G scheduler design: From theoretical framework to implementation,

    Z. Gu, C. She, W. Hardjawana, S. Lumb, D. McKechnie, T. Essery, and B. Vucetic, “Knowledge-assisted deep reinforcement learning in 5G scheduler design: From theoretical framework to implementation,” IEEE Journal on Selected Areas in Communications , vol. 39, no. 7, pp. 2014– 2...

    2014

  5. [12]

    DetFed: Dynamic resource scheduling for deterministic federated learning over time-sensitive networks,

    D. Yang, W. Zhang, Q. Ye, C. Zhang, N. Zhang, C. Huang, H. Zhang, and X. Shen, “DetFed: Dynamic resource scheduling for deterministic federated learning over time-sensitive networks,” IEEE Transactions on Mobile Computing, 2023

    2023

  6. [13]

    Predictable wireless networked scheduling for bridging hybrid time-sensitive and real-time services,

    Y . Zhou, L. Feng, X. Jiang, W. Li, and F. Zhou, “Predictable wireless networked scheduling for bridging hybrid time-sensitive and real-time services,” IEEE Transactions on Communications , 2024

    2024

  7. [14]

    Multicell OFDMA downlink resource allocation using a graphic framework,

    R. Y . Chang, Z. Tao, J. Zhang, and C.-C. J. Kuo, “Multicell OFDMA downlink resource allocation using a graphic framework,” IEEE Trans- actions on Vehicular Technology, vol. 58, no. 7, pp. 3494–3507, 2009

    2009

  8. [15]

    Graph-based resource sharing in vehicular communication,

    L. Liang, S. Xie, G. Y . Li, Z. Ding, and X. Yu, “Graph-based resource sharing in vehicular communication,” IEEE Transactions on Wireless Communications, vol. 17, no. 7, pp. 4579–4592, 2018

    2018

  9. [16]

    Minimum interference channel assignment in multiradio wireless mesh networks,

    A. P. Subramanian, H. Gupta, S. R. Das, and J. Cao, “Minimum interference channel assignment in multiradio wireless mesh networks,” IEEE transactions on mobile computing , vol. 7, no. 12, pp. 1459–1473, 2008

    2008

  10. [17]

    Graph representation learning for contention and interference management in wireless networks,

    Z. Gu, B. Vucetic, K. Chikkam, P. Aliberti, and W. Hardjawana, “Graph representation learning for contention and interference management in wireless networks,” IEEE/ACM Transactions on Networking , pp. 1–16, 2024

    2024

  11. [18]

    Improved approximation algo- rithms for maximum cut and satisfiability problems using semidefinite programming,

    M. X. Goemans and D. P. Williamson, “Improved approximation algo- rithms for maximum cut and satisfiability problems using semidefinite programming,” Journal of the ACM (JACM) , vol. 42, no. 6, pp. 1115– 1145, 1995

    1995

  12. [19]

    Large-scale convex optimization for dense wireless cooperative networks,

    Y . Shi, J. Zhang, B. O’Donoghue, and K. B. Letaief, “Large-scale convex optimization for dense wireless cooperative networks,” IEEE Transactions on Signal Processing, vol. 63, no. 18, pp. 4729–4743, 2015

    2015

  13. [20]

    Stability properties of constrained queueing systems and scheduling policies for maximum throughput in multihop radio networks,

    L. Tassiulas and A. Ephremides, “Stability properties of constrained queueing systems and scheduling policies for maximum throughput in multihop radio networks,” in 29th IEEE Conference on Decision and Control. IEEE, 1990, pp. 2130–2132

    1990

  14. [21]

    Understanding the capacity region of the greedy maximal scheduling algorithm in multihop wireless networks,

    C. Joo, X. Lin, and N. B. Shroff, “Understanding the capacity region of the greedy maximal scheduling algorithm in multihop wireless networks,” IEEE/ACM transactions on networking , vol. 17, no. 4, pp. 1132–1145, 2009

    2009

  15. [22]

    Analyzing the performance of greedy maximal scheduling via local pooling and graph theory,

    B. Birand, M. Chudnovsky, B. Ries, P. Seymour, G. Zussman, and Y . Zwols, “Analyzing the performance of greedy maximal scheduling via local pooling and graph theory,” IEEE/ACM Transactions on Net- working, vol. 20, no. 1, pp. 163–176, 2011

    2011

  16. [23]

    On the power of randomization for scheduling real-time traffic in wireless networks,

    C. Tsanikidis and J. Ghaderi, “On the power of randomization for scheduling real-time traffic in wireless networks,” IEEE/ACM Trans- actions on Networking , vol. 29, no. 4, pp. 1703–1716, 2021

    2021

  17. [24]

    Opportunistic scheduling using statistical information of wireless channels,

    Z. Gu, W. Hardjawana, and B. Vucetic, “Opportunistic scheduling using statistical information of wireless channels,” IEEE Transactions on Wireless Communications, 2024

    2024

  18. [25]

    A survey of recent advances in optimization methods for wireless communications,

    Y .-F. Liu, T.-H. Chang, M. Hong, Z. Wu, A. M.-C. So, E. A. Jorswieck, and W. Yu, “A survey of recent advances in optimization methods for wireless communications,” accepted in IEEE Journal on Selected Areas in Communications, 2024

    2024

  19. [26]

    SDPT3—a MATLAB soft- ware package for semidefinite programming, version 1.3,

    K.-C. Toh, M. J. Todd, and R. H. T ¨ut¨unc¨u, “SDPT3—a MATLAB soft- ware package for semidefinite programming, version 1.3,” Optimization methods and software , vol. 11, no. 1-4, pp. 545–581, 1999

    1999

  20. [27]

    Implementation and evaluation of SDPA 6.0 (semidefinite programming algorithm 6.0),

    M. Yamashita, K. Fujisawa, and M. Kojima, “Implementation and evaluation of SDPA 6.0 (semidefinite programming algorithm 6.0),” Optimization Methods and Software , vol. 18, no. 4, pp. 491–505, 2003

    2003

  21. [28]

    Conic optimization via operator splitting and homogeneous self-dual embedding,

    B. O’donoghue, E. Chu, N. Parikh, and S. Boyd, “Conic optimization via operator splitting and homogeneous self-dual embedding,” Journal of Optimization Theory and Applications , vol. 169, pp. 1042–1068, 2016

    2016

  22. [29]

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

    2004

  23. [30]

    Strang, Linear algebra and its applications

    G. Strang, Linear algebra and its applications. Belmont, CA: Thomson, Brooks/Cole, 2006

    2006

  24. [31]

    A combinatorial, primal-dual approach to semidefinite programs,

    S. Arora and S. Kale, “A combinatorial, primal-dual approach to semidefinite programs,” in Proceedings of the thirty-ninth annual ACM symposium on Theory of computing , 2007, pp. 227–236

    2007

  25. [32]

    Fast SDP algorithms for constraint satisfaction problems,

    D. Steurer, “Fast SDP algorithms for constraint satisfaction problems,” in Proceedings of the twenty-first annual ACM-SIAM symposium on Discrete Algorithms. SIAM, 2010, pp. 684–697

    2010

  26. [34]

    Principle and performance of semi-persistent scheduling for V oIP in LTE system,

    D. Jiang, H. Wang, E. Malkamaki, and E. Tuomaala, “Principle and performance of semi-persistent scheduling for V oIP in LTE system,” in 2007 International Conference on Wireless Communications, Network- ing and Mobile Computing . IEEE, 2007, pp. 2861–2864

    2007

  27. [35]

    Quasi-static multiple- antenna fading channels at finite blocklength,

    W. Yang, G. Durisi, T. Koch, and Y . Polyanskiy, “Quasi-static multiple- antenna fading channels at finite blocklength,” IEEE Transactions on Information Theory, vol. 60, no. 7, pp. 4232–4265, 2014

    2014

  28. [36]

    D. B. West et al. , Introduction to graph theory . Prentice hall Upper Saddle River, 2001, vol. 2

    2001

  29. [37]

    On eigenvalues and colorings of graphs,

    A. J. Hoffman, “On eigenvalues and colorings of graphs,” in Selected Papers Of Alan J Hoffman: With Commentary . World Scientific, 2003, pp. 407–419

    2003

  30. [38]

    A faster interior point method for semidefinite programming,

    H. Jiang, T. Kathuria, Y . T. Lee, S. Padmanabhan, and Z. Song, “A faster interior point method for semidefinite programming,” in 2020 IEEE 61st annual symposium on foundations of computer science (FOCS) . IEEE, 2020, pp. 910–918

    2020

  31. [39]

    Large-scale convex optimization for ultra-dense cloud-ran,

    Y . Shi, J. Zhang, K. B. Letaief, B. Bai, and W. Chen, “Large-scale convex optimization for ultra-dense cloud-ran,” IEEE Wireless Communications, vol. 22, no. 3, pp. 84–91, 2015

    2015

  32. [40]

    Improved approximation algorithms for max k-cut and max bisection,

    A. Frieze and M. Jerrum, “Improved approximation algorithms for max k-cut and max bisection,” Algorithmica, vol. 18, no. 1, pp. 67–81, 1997

    1997

  33. [41]

    R. B. Lehoucq, D. C. Sorensen, and C. Yang, ARPACK users’ guide: solution of large-scale eigenvalue problems with implicitly restarted Arnoldi methods. SIAM, 1998

    1998

  34. [42]

    A decision-theoretic generalization of on-line learning and an application to boosting,

    Y . Freund and R. E. Schapire, “A decision-theoretic generalization of on-line learning and an application to boosting,” Journal of computer and system sciences , vol. 55, no. 1, pp. 119–139, 1997

    1997

  35. [43]

    A new scaling and squaring algorithm for the matrix exponential,

    A. H. Al-Mohy and N. J. Higham, “A new scaling and squaring algorithm for the matrix exponential,” SIAM Journal on Matrix Analysis and Applications, vol. 31, no. 3, pp. 970–989, 2010

    2010

  36. [44]

    Study on NR Industrial Internet of Things (IoT),

    3GPP, “Study on NR Industrial Internet of Things (IoT),” 3GPP, TS 38.825, 2019, v16.0.0

    2019

  37. [45]

    Propagation data and prediction methods for the planning of indoor radiocommunication systems and radio local area networks in the frequency range 300 MHz to 100 GHz,

    P. Series, “Propagation data and prediction methods for the planning of indoor radiocommunication systems and radio local area networks in the frequency range 300 MHz to 100 GHz,” Recommendation ITU-R , pp. 1238–8, 2015

    2015

  38. [46]

    SciPy 1.0: Fundamental Algorithms for Scientific Computing in Python,

    P. Virtanen et al. , “SciPy 1.0: Fundamental Algorithms for Scientific Computing in Python,” Nature Methods, vol. 17, pp. 261–272, 2020

    2020

  39. [47]

    Reducing the bandwidth of sparse symmetric matrices,

    E. Cuthill and J. McKee, “Reducing the bandwidth of sparse symmetric matrices,” in Proceedings of the 1969 24th national conference , 1969, pp. 157–172

    1969

  40. [48]

    Penrose, Random geometric graphs

    M. Penrose, Random geometric graphs. Oxford University Press, 2003, vol. 5

    2003

  41. [49]

    MATLAB, https://mathworks.com/help/symbolic/eigenvalues.html

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.