Asynchronous Distributed Voltage Control in Active Distribution Networks

Boyu Qin; Feng Liu; Yifan Su; Zhaojian Wang

REVIEW 4 major objections 4 minor 31 references

Asynchronous Distributed Voltage Control in Active Distribution Networks

T0 review · 4 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read An asynchronous distributed controller built on a partial primal-dual gradient method converges almost surely to the optimal voltage profile even with bounded delays and non-identical update rates.

desk verdict A solid asynchronous distributed voltage-control paper whose proof only covers homogeneous feeders, while the headline 123-bus test is non-homogeneous and fixes K=1 by hand; the gap is real but the theoretical core is still worth refereeing. read the letter →

arxiv 1908.06548 v1 pith:FFVCVWZ3 submitted 2019-08-19 eess.SY cs.SY

Zhaojian Wang , Feng Liu , Yifan Su , Boyu Qin This is my paper · ORCID

classification eess.SYcs.SY MSC 93A1490C2568W15

keywords asynchronousdistributedcontrolvoltagepartialprimal-dualgradientalgorithmoperatorsplittingrandomcommunicationdelaysdistributionnetworksenergyresourcesfixed-pointiteration

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

This paper claims that voltage regulation in a radial distribution feeder can be carried out by an asynchronous distributed controller that lets each bus update on its own local clock instead of waiting for the slowest neighbor. The controller is built on a partial primal-dual gradient method and adjusts both active and reactive power of distributed energy resources using only neighbor and two-hop-neighbor communication. Under convex costs with Lipschitz gradients, bounded communication delays, and a homogeneity condition on line impedances, the paper proves that the iterates converge with probability 1 to a point satisfying the KKT conditions of the optimal voltage-control problem. The payoff is practical: real communication links have random delays and devices sample at different rates, and a synchronous scheme is only as fast as its slowest component.

What carries the argument

The load-bearing object is the averaged fixed-point operator $S = S_1 S_2$ produced by splitting the primal-dual optimality conditions. With $C(w) = (F(z), \vartheta_a + B^2\lambda)$ and $D(w) = (N_\Omega(z) - \mathrm{col}(K\lambda, \lambda), (K I_n, I_n)z)$, the synchronous step is $\tilde{w} = (\mathrm{Id} + \Gamma^{-1}D)^{-1}(\mathrm{Id} - \Gamma^{-1}C)w$ and $w_{t+1} = w_t + \eta(\tilde{w}_t - w_t)$, where $\Gamma$ is the positive-definite block matrix of stepsizes $\alpha_{pq}$ and $\alpha_\lambda$. Because $C$ is cocoercive and $D$ is maximally monotone, the resolvent $S_1$ is firmly nonexpansive and $S_2$ is averaged, making $S$ an averaged operator; the synchronous updates are a Krasnosel'skii-Mann iteration, a standard fixed-point scheme for such operators, and the asynchronous version is the same iteration with random coordinate selection and delayed arguments, which falls under an existing almost-sure convergence theorem. The homogeneous-network identity $R = KX$, which turns the power-flow constraint into $BV = Kp + q + \vartheta_s$, is what lets the operators separate cleanly.

What would settle it

Run ASDVC on a small non-homogeneous feeder, say three buses with $r/x$ ratios 0.4, 1.2, and 2.0, with a convex quadratic cost, bounded delays, and stepsizes satisfying the theorem, and compare the terminal point with the KKT conditions of the true non-homogeneous optimal power-flow problem. If the iterates converge to a point violating those KKT conditions, or fail to converge, then the homogeneity assumption is doing essential work and the claimed tolerance of heterogeneous feeders is unsupported.

Watch

Extended reading notes

Core claim

The central claim is that an asynchronous version of the partial primal-dual gradient algorithm converges almost surely to the primal-dual optimal solution of a linearized optimal voltage-control problem. Theorem 3 shows that any equilibrium of the ASDVC iteration satisfies the KKT conditions of the optimization problem, hence is primal-dual optimal; Theorem 4 shows that, under bounded delays, uniform random bus activation, and an explicit stepsize bound involving the delay bound, the curvature of the cost, and the spectrum of the network matrix $B$, the sequence $w_t$ converges with probability 1 to that equilibrium. The proof rewrites the synchronous update as a fixed-point iteration with an averaged operator via operator splitting, and then views the asynchronous updates as a randomized block-coordinate fixed-point iteration with delayed information. The framework unifies two practically distinct sources of asynchrony: random communication delays and non-identical sampling or control rates.

Load-bearing premise

The proof assumes the feeder is homogeneous in the sense that every line has the same resistance-to-reactance ratio $K$, so the voltage constraint simplifies to $BV = Kp + q + \vartheta_s$; if real line ratios vary widely, that constraint is only approximate and the convergence guarantee is not formally in force.

Editorial extensions

If this is right

Buses can update on their own local clocks; the slowest bus or congested channel no longer dictates the iteration rate, so the controller can track time-varying load and solar profiles better than synchronous schemes under delay.
The same convergence proof covers both random communication delays and non-identical sampling or control rates, because both are represented as bounded intervals between local iterations.
Only neighbor and two-hop-neighbor communication is needed; no central coordinator or global model is required for the iteration itself.
The objective function only needs convexity with a Lipschitzian gradient, not strong convexity, so general DER cost terms fit the framework.
An online implementation using local voltage and power measurements makes the controller adjustable to time-varying operating conditions.

Reading between the lines

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

The homogeneity assumption could likely be relaxed to a bounded-heterogeneity condition: if $R$ and $X$ differ by a small perturbation, one would expect convergence to a neighborhood of the KKT point whose radius scales with the perturbation; the paper does not prove this.
The proof route suggests a rate prediction: convergence speed should degrade as the delay bound $\chi$ grows and as the spectrum of $B$ spreads through the parameter $\kappa$, so feeders with strongly heterogeneous $r/x$ should show slower convergence; this is testable in simulation.
If the online implementation's voltage measurements are noisy, the iteration becomes a stochastic approximation rather than a deterministic fixed-point method; whether the almost-sure guarantee survives measurement noise is an open question not addressed here.
Extending the same operator-splitting argument to unbalanced three-phase feeders would require a different matrix $B$ and would likely break the clean two-hop sparsity; the paper lists three-phase symmetry as a limitation.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

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

Two things to know. The paper genuinely extends distributed voltage control into asynchronous territory—delayed and non-uniform updates for both P and Q, with a convergence proof via operator splitting and ARock. That is not a trivial patch. But the main numerical claim rests on an assumption the paper's own headline test violates: the theory assumes a homogeneous feeder (constant R/X ratio), and the IEEE-123 system is not homogeneous, with R/X ranging from 0.42 to 2.02. Setting K=1 by hand and calling it robustness is not supported by any bound.

Let me give credit where it's due. The fixed-point reformulation is clean and the operator properties are checked carefully. The 8-bus test matches the theory (K=2), and the convergence behavior is sensible. The online implementation using local voltage measurements to estimate the offset term is a practical touch. The use of ARock is appropriate: Theorem 4 is a direct application of Peng et al., and the conditions are stated explicitly. The citation pattern is fine.

The soft spots, in proportion. The homogeneity issue is load-bearing. Lemma 2 and Theorem 3 both rely on R = KX and B R = K I to build the averaged operator S. Without a constant K, equation (5b) is not equivalent to the actual linearized DistFlow, so converging to a KKT point of the modified problem says nothing about the real feeder. The authors acknowledge eliminating the homogeneity restriction is ongoing work, but that does not fix the gap between the theorem and the 123-bus experiment. This is not a small technicality; it is the paper's practical core. Second, the ARock hypotheses are imported without close verification in the network context. The uniform activation assumption in the global clock is stated but not justified for bus-local clocks; bounded delay is assumed but not derived from the two-hop communication scheme. Third, the stepsize conditions depend on beta, kappa, sigma_max, and vartheta, but there is no guidance on choosing them beyond existence. These are not fatal to the theoretical contribution, but they do limit how much an operator can trust the paper's claims.

Who gets value: researchers working on asynchronous distributed optimization in power systems. The paper is a credible building block, not a paradigm shift. It deserves a serious referee: the math is mostly sound, the problem is well motivated, and the gap is identifiable and fixable in revision—either by proving a robust version or by repositioning the claims. I would send it to review, with a clear request to address the homogeneity gap.

Referee Report

4 major / 4 minor

Summary. This paper proposes an asynchronous distributed voltage control scheme (ASDVC) for radial distribution networks. The authors formulate an optimal voltage control problem (5) under a homogeneous-network assumption, use a partial primal-dual gradient algorithm with local clocks, and convert the iteration into a fixed-point form via operator splitting. They prove that any equilibrium satisfies the KKT conditions (Theorem 3) and that the asynchronous iteration converges almost surely under bounded delays and a stepsize condition (Theorem 4). The paper also provides an online implementation formula and validates the controller on an 8-bus feeder and the IEEE-123 system using real load and solar data.

Significance. The contribution is potentially useful: it unifies active and reactive power control in an asynchronous distributed framework, avoids strong convexity assumptions, and builds on standard monotone operator theory. The operator-splitting reformulation is clean and the averaging argument is largely standard. However, the practical significance is limited by the homogeneity assumption in all theoretical results; the main scalability experiment violates this assumption, and the claimed robustness is unsupported. In addition, the transfer of the ARock theorem to the weighted-norm setting is not justified, and the random activation model is not stated as a hypothesis of Theorem 4. With these gaps addressed, the manuscript could be a solid contribution to asynchronous distributed voltage control.

major comments (4)

[Section 2.2, Section 7.2, Conclusion] The theoretical results (Lemmas 1–2 and Theorems 3–4) depend on the homogeneous-network assumption: Eq. (5b) is derived from R = KX in Section 2.2. The IEEE-123 system is not homogeneous (r/x ranges from 0.42 to 2.02), and the simulation sets K = 1 by hand. The claim in Section 7.2 that this 'shows the robustness of our method' is not backed by any theorem or quantitative error bound. This is load-bearing because the practical contribution rests on the non-homogeneous validation. Please either restrict the claims, provide a perturbation/error analysis for non-homogeneous networks, or give a separate theoretical treatment of the heterogeneous case.
[Section 5.3, Lemma 2, Theorem 4] The convergence proof applies the ARock theorem (Peng et al. 2016, Lemma 13 and Theorem 14) to iteration (30). However, Lemma 2 only shows that the composed operator S, and hence T, is averaged with respect to the Γ-induced norm, while ARock operates in the Euclidean norm and the block-coordinate update (27) uses the Euclidean Hadamard product. The passage from a Γ-nonexpansive operator to the ARock assumptions is not justified. Please show that ARock applies under the Γ-norm geometry, or reformulate the iteration so that the Euclidean assumptions are satisfied.
[Section 5.3, Theorem 4] Theorem 4 lists only Assumptions 1 and 2, but its proof relies on the ARock result that requires random coordinate selection with positive probability. The informal global-clock construction assumes a uniform activation distribution, but this is not stated as a hypothesis of Theorem 4, nor is it enforced by Algorithm 1. Moreover, Assumption 2 bounds τ^{t_j}_j in local-clock units, whereas the ARock delay bound χ refers to global iterations; the mapping between these quantities is not defined. The theorem statement should include the random activation assumption and a precise global delay bound.
[Section 5.2, Theorem 3] The theorem asserts that the KKT point is 'the primal-dual optimal solution' to problem (5). The paper cites Ruszczynski Theorem 3.25 for KKT conditions but does not verify a constraint qualification (e.g., Slater's condition) for the apparent-power constraint (5e). Since KKT conditions are necessary, not sufficient in general, the optimality conclusion requires either a Slater-point check or an explicit statement of the qualification assumption.

minor comments (4)

[Section 6.2, Eq. (31)] Eq. (31) uses V^m_k without stating that it denotes the measured value of U^2_k/2; earlier V is defined as U^2/2, so the notation should be clarified to avoid confusion.
[Section 7, Figs. 3 and 6] The horizontal axis label 'Number of average iterations' is not defined; please specify how the average is taken across buses. Also, in Fig. 6 the shaded curves and legends are difficult to distinguish; consider using distinct line styles.
[Section 6.1] The statement that 'the topology of a distribution network does not change frequently' should be made precise: the online implementation assumes the matrices B and B^2 are fixed, and this assumption should be stated explicitly.
[Throughout] There are several typographical issues, including the title 'V oltage' and inconsistent spelling of 'different/diﬀerent'; the paper would benefit from a careful proofreading pass.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the convergence and optimality derivations are self-contained, and the homogeneity caveat is an assumption-validity issue, not a circular step.

full rationale

Walking the derivation chain, the paper starts from the optimal voltage control problem (5) under the explicit homogeneous-network assumption in Section 2.2, which turns the linearized DistFlow relation into BV = Kp + q + varpi_s in Eq. (5b). The ASDVC updates in (8) are a partial primal-dual gradient iteration, and the reformulation in Eqs. (9)-(12) and the operator definitions in (11) are algebraic rearrangements of those updates. Lemma 1 and Lemma 2 invoke standard monotone operator theory (Bauschke et al. 2011, Combettes and Yamada 2015, Yi and Pavel 2019b) to show that the fixed-point operator S is averaged; these results are external and parameter-free, not equivalent to the paper's target claim. Theorem 3 compares the equilibrium conditions (26) with the KKT conditions (25) of the stated optimization problem; since (25) is the standard necessary optimality condition for (5), this is a direct equivalence check, not an imported conclusion. Theorem 4 maps ASDVC to the delayed randomized block-coordinate iteration (28)-(30) and invokes the external ARock convergence theorem of Peng et al. (2016); no fitted parameter is later renamed as a prediction, and the stepsize conditions are stated as assumptions rather than calibrated to the simulation results. The paper does contain a genuine limitation: Section 7.2 applies the algorithm to the IEEE-123 system, noting that it is not homogeneous with r_ij/x_ij from 0.42 to 2.02, and then sets K = 1 by hand to claim robustness, even though the theoretical guarantees in Theorems 3-4 are proved only for homogeneous networks. Section 8 explicitly concedes that eliminating the homogeneity restriction is ongoing work. This is a theory-experiment mismatch and a robustness gap, but it is not circularity: the simulations do not provide the proof of the theorem, and no equation in the derivation is defined in terms of the result it purports to establish. The self-citations, such as Wang et al. (2019) in Section 4, are contextual references to related primal-dual formulations and are not load-bearing for the convergence proof. Overall, no circular step was identified.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The central convergence proof rests on four domain assumptions (convexity/Lipschitz, bounded delays, uniform activation, homogeneity) plus an external theorem (ARock) that is cited but not verified in detail for this operator. No new physical entities are introduced; the method is algorithmic.

free parameters (2)

K (resistance-to-reactance ratio) = 1 (chosen in IEEE-123 simulation; true network ratio varies 0.42 to 2.02)
In the theory K is the constant line ratio required for homogeneity; in the heterogeneous IEEE-123 simulation K=1 is set by hand, so the simulated system does not satisfy the model assumption.
Stepsizes alpha_pq, alpha_lambda, eta = not fitted to data; set to satisfy the theorem's inequalities
These algorithm parameters are chosen by hand to satisfy Gamma - kappa I positive semidefinite and the eta bound in Theorem 4. The central claim depends on them, but they are not tuned to the simulation results.

assumptions (5)

domain assumption The functions g_i are convex and have vartheta-Lipschitz gradients (Assumption 1).
This is required for the cocoercivity of operator C in Lemma 1 and is stated in Section 3.
domain assumption The maximal delay between consecutive iterations is bounded by chi (Assumption 2).
Stated in Section 5.3; the ARock convergence theorem requires bounded delays.
domain assumption Bus activation in the global clock follows a uniform distribution (Section 5.3).
Invoked to fit the algorithm into the randomized block-coordinate ARock framework; the actual activation process in practice may not be uniform.
domain assumption The distribution network is homogeneous, R = KX (Section 2.2).
Required for the linearized model BV = Kp + q + varpi_s and for the subsequent KKT and convergence analysis.
standard math The ARock convergence theorem (Peng et al. 2016, Lemmas 13 and Theorem 14) applies to the constructed nonexpansive operator T and the delayed coordinate update (28).
The proof of Theorem 4 delegates the convergence guarantee to this external result without fully verifying all of its assumptions for this specific operator and delay model.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Asynchronous Distributed Voltage Control in Active Distribution Networks." pith.science (2026). https://pith.science/paper/FFVCVWZ3

@misc{pith2026190806548,
  author       = {Pith},
  title        = {Pith review of: Asynchronous Distributed Voltage Control in Active Distribution Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FFVCVWZ3}},
  note         = {Machine review of arXiv:1908.06548}
}

read the original abstract

With the explosion of distributed energy resources (DERs), voltage regulation in distribution networks has been facing a great challenge. This paper derives an asynchronous distributed voltage control strategy based on the partial primal-dual gradient algorithm, where both active and reactive controllable power of DERs are considered. Different types of asynchrony due to imperfect communication or practical limits, such as random time delays and non-identical sampling/control rates, are fitted into a unified analytic framework. The asynchronous algorithm is then converted into a fixed-point problem by employing the operator splitting method, which leads to a convergence proof with mild conditions. Moreover, an online implementation method is provided to make the controller adjustable to time-varying environments. Finally, numerical experiments are carried out on a rudimentary 8-bus system and the IEEE-123 distribution network to verify the effectiveness of the proposed method.

Figures

Figures reproduced from arXiv: 1908.06548 by the authors.

**Figure 1.** Two-step communications neighbors’ information is needed to obtain B˜ jk and λk,tk−τ tk k , k ∈ N 2 j . For B˜ jk, it can be obtained from twice neighboring communications. In addition, as the topology of a distribution network does not change frequently, B˜ jk can be obtained in advance. For λk,tk−τ tk k , k ∈ N 2 j , it also can be obtained from neighboring communications, which is illustrated in Fig.1. Node i c… view at source ↗

**Figure 2.** The graph of the 8-bus distribution network kW [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗

**Figure 4.** IEEE 123-bus system 0 5 10 15 20 Time (Hour) 0 1 2 3 4 5 6 Active Power Reactive Power Active/Reactive Power (kW/kVar) 24 0 5 10 15 20 Time (Hour) 0 1 2 3 4 5 6 Active Power Reactive Power Active/Reactive Power (kW/kVar) 24 Solar Generation 0 5 10 15 20 25 30 35 40 45 50 0 5 10 15 20 Time (Hour) 24 Active Power Reactive Power Solar Generation Active/Reactive Power (kW/kVar) 0 5 10 15 20 25 30 35 40 45 50 0 5 10 15 2… view at source ↗

Figures from the paper (2 more)

**Figure 5.** Figure 5: Active, reactive loads and solar generation within 24 hours [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]

**Figure 6.** Figure 6: Daily voltage mismatch error with SDVC and ASDVC under ran [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

31 extracted references · 30 canonical work pages

[1]

, author Kouveliotis-Lysikatos, I.N

author Antoniadou-Plytaria, K.E. , author Kouveliotis-Lysikatos, I.N. , author Georgilakis, P.S. , author Hatziargyriou, N.D. , year 2017 . title Distributed and decentralized voltage control of smart distribution networks: models, methods, and future research . journal IEEE Trans. Smart Grid volume 8 , pages 2999--3008

work page 2017
[2]

, author Wu, F.F

author Baran, M. , author Wu, F.F. , year 1989 a. title Optimal sizing of capacitors placed on a radial distribution system . journal IEEE Trans. power Delivery volume 4 , pages 735--743

work page 1989
[3]

, author Wu, F.F

author Baran, M.E. , author Wu, F.F. , year 1989 b. title Optimal capacitor placement on radial distribution systems . journal IEEE Trans. power Delivery volume 4 , pages 725--734

work page 1989
[4]

, author Combettes, P.L

author Bauschke, H.H. , author Combettes, P.L. , et al., year 2011 . title Convex analysis and monotone operator theory in Hilbert spaces . volume volume 408 . publisher Springer

work page 2011
[5]

, author Carli, R

author Bolognani, S. , author Carli, R. , author Cavraro, G. , author Zampieri, S. , year 2015 . title Distributed reactive power feedback control for voltage regulation and loss minimization . journal IEEE Trans. Autom. Control volume 60 , pages 966--981

work page 2015
[6]

, author Yamada, I

author Combettes, P.L. , author Yamada, I. , year 2015 . title Compositions and convex combinations of averaged nonexpansive operators . journal Journal of Mathematical Analysis and Applications volume 425 , pages 55--70

work page 2015
[7]

, author Clarke, C.R

author Farivar, M. , author Clarke, C.R. , author Low, S.H. , author Chandy, K.M. , year 2011 . title Inverter var control for distribution systems with renewables , in: booktitle Smart Grid Communications (SmartGridComm), 2011 IEEE International Conference on , organization IEEE . pp. pages 457--462

work page 2011
[8]

, author Neal, R

author Farivar, M. , author Neal, R. , author Clarke, C. , author Low, S. , year 2012 . title Optimal inverter var control in distribution systems with high pv penetration , in: booktitle Power and Energy Society General Meeting, 2012 IEEE , organization IEEE . pp. pages 1--7

work page 2012

Show all 31 references

[9]

, author Nedi \'c , A

author Hale, M.T. , author Nedi \'c , A. , author Egerstedt, M. , year 2017 . title Asynchronous multiagent primal-dual optimization . journal IEEE Trans. Automatic Control volume 62 , pages 4421--4435

2017
[10]

, author Chen, L

author Han, Y. , author Chen, L. , author Ma, H. , author Wang, Z. , year 2014 . title Optimization of reactive power compensation for distribution power system with small hydro power , in: booktitle Power System Technology (POWERCON), 2014 International Conference on , organi...

2014
[11]

, author Wang, G

author Kekatos, V. , author Wang, G. , author Conejo, A.J. , author Giannakis, G.B. , year 2015 a. title Stochastic reactive power management in microgrids with renewables . journal IEEE Trans. Power Syst. volume 30 , pages 3386--3395

2015
[12]

, author Zhang, L

author Kekatos, V. , author Zhang, L. , author Giannakis, G.B. , author Baldick, R. , year 2015 b. title Fast localized voltage regulation in single-phase distribution grids , in: booktitle Smart Grid Communications (SmartGridComm), 2015 IEEE International Conference on , orga...

2015
[13]

, author Zhang, L

author Kekatos, V. , author Zhang, L. , author Giannakis, G.B. , author Baldick, R. , year 2016 . title Voltage regulation algorithms for multiphase power distribution grids . journal IEEE Trans. Power Syst volume 31 , pages 3913--3923

2016
[14]

, author Zhao, C

author Li, N. , author Zhao, C. , author Chen, L. , year 2016 . title Connecting automatic generation control and economic dispatch from an optimization view . journal IEEE Trans. Control Netw. Syst. volume 3 , pages 254--264

2016
[15]

, author Shi, W

author Liu, H.J. , author Shi, W. , author Zhu, H. , year 2017 . title Decentralized dynamic optimization for power network voltage control . journal IEEE Trans. Signal Inf. Process. Networks volume 3 , pages 568--579

2017
[16]

, author Shi, W

author Liu, H.J. , author Shi, W. , author Zhu, H. , year 2018 a. title Distributed voltage control in distribution networks: Online and robust implementations . journal IEEE Trans. Smart Grid volume 9 , pages 6106--6117

2018
[17]

, author Shi, W

author Liu, H.J. , author Shi, W. , author Zhu, H. , year 2018 b. title Hybrid voltage control in distribution networks under limited communication rates . journal IEEE Trans. Smart Grid

2018
[18]

, year 2000

author Meyer, C.D. , year 2000 . title Matrix analysis and applied linear algebra . volume volume 71 . publisher Siam

2000
[19]

title Solrmap utah geological survey

author NREL , year 2018 . title Solrmap utah geological survey . howpublished https://midcdmz.nrel.gov/usep_cedar/

2018
[20]

, author Xu, Y

author Peng, Z. , author Xu, Y. , author Yan, M. , author Yin, W. , year 2016 . title Arock: an algorithmic framework for asynchronous parallel coordinate updates . journal SIAM J. Sci. Comput. volume 38 , pages A2851--A2879

2016
[21]

, author Ruszczynski, A

author Ruszczy \'n ski, A.P. , author Ruszczynski, A. , year 2006 . title Nonlinear optimization . volume volume 13 . publisher Princeton university press

2006
[22]

, author Backhaus, S

author S ulc, P. , author Backhaus, S. , author Chertkov, M. , year 2014 . title Optimal distributed control of reactive power via the alternating direction method of multipliers . journal IEEE Trans. Energy Convers. volume 29 , pages 968--977

2014
[23]

, author Hill, D.J

author Tang, Z. , author Hill, D.J. , author Liu, T. , year 2019 . title Fast distributed reactive power control for voltage regulation in distribution networks . journal IEEE Trans. Power Syst. volume 34 , pages 802--805

2019
[24]

, author Sulc, P

author Turitsyn, K. , author Sulc, P. , author Backhaus, S. , author Chertkov, M. , year 2011 . title Options for control of reactive power by distributed photovoltaic generators . journal Proceedings of the IEEE volume 99 , pages 1063--1073

2011
[25]

title Individual household electric power consumption data set

author UCI , year 2012 . title Individual household electric power consumption data set . howpublished https://archive.ics.uci.edu/ml/datasets/individual+household+electric+power+consumption

2012
[26]

, author Liu, F

author Wang, Z. , author Liu, F. , author Low, S.H. , author Zhao, C. , author Mei, S. , year 2019 . title Distributed frequency control with operational constraints, part ii: Network power balance . journal IEEE Trans. Smart Grid volume 10 , pages 53--64

2019
[27]

, author Pavel, L

author Yi, P. , author Pavel, L. , year 2019 a. title Asynchronous distributed algorithms for seeking generalized nash equilibria under full and partial decision information . journal IEEE Trans. Cybern., DOI, 10.1109/TCYB.2019.2908091

2019
[28]

, author Pavel, L

author Yi, P. , author Pavel, L. , year 2019 b. title An operator splitting approach for distributed generalized nash equilibria computation . journal Automatica volume 102 , pages 111--121

2019
[29]

, author Lam, A.Y

author Zhang, B. , author Lam, A.Y. , author Dom \' nguez-Garc \' a, A.D. , author Tse, D. , year 2015 . title An optimal and distributed method for voltage regulation in power distribution systems . journal IEEE Trans. Power Syst. volume 30 , pages 1714--1726

2015
[30]

, author Chen, L

author Zhou, X. , author Chen, L. , author Farivar, M. , author Liu, Z. , author Low, S. , year 2018 . title Reverse and forward engineering of local voltage control in distribution networks . journal arXiv preprint arXiv:1801.02015

2018 arXiv
[31]

, author Liu, H.J

author Zhu, H. , author Liu, H.J. , year 2016 . title Fast local voltage control under limited reactive power: Optimality and stability analysis . journal IEEE Trans. Power Syst. volume 31 , pages 3794--3803

2016

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

, author Kouveliotis-Lysikatos, I.N

author Antoniadou-Plytaria, K.E. , author Kouveliotis-Lysikatos, I.N. , author Georgilakis, P.S. , author Hatziargyriou, N.D. , year 2017 . title Distributed and decentralized voltage control of smart distribution networks: models, methods, and future research . journal IEEE Trans. Smart Grid volume 8 , pages 2999--3008

work page 2017

[2] [2]

, author Wu, F.F

author Baran, M. , author Wu, F.F. , year 1989 a. title Optimal sizing of capacitors placed on a radial distribution system . journal IEEE Trans. power Delivery volume 4 , pages 735--743

work page 1989

[3] [3]

, author Wu, F.F

author Baran, M.E. , author Wu, F.F. , year 1989 b. title Optimal capacitor placement on radial distribution systems . journal IEEE Trans. power Delivery volume 4 , pages 725--734

work page 1989

[4] [4]

, author Combettes, P.L

author Bauschke, H.H. , author Combettes, P.L. , et al., year 2011 . title Convex analysis and monotone operator theory in Hilbert spaces . volume volume 408 . publisher Springer

work page 2011

[5] [5]

, author Carli, R

author Bolognani, S. , author Carli, R. , author Cavraro, G. , author Zampieri, S. , year 2015 . title Distributed reactive power feedback control for voltage regulation and loss minimization . journal IEEE Trans. Autom. Control volume 60 , pages 966--981

work page 2015

[6] [6]

, author Yamada, I

author Combettes, P.L. , author Yamada, I. , year 2015 . title Compositions and convex combinations of averaged nonexpansive operators . journal Journal of Mathematical Analysis and Applications volume 425 , pages 55--70

work page 2015

[7] [7]

, author Clarke, C.R

author Farivar, M. , author Clarke, C.R. , author Low, S.H. , author Chandy, K.M. , year 2011 . title Inverter var control for distribution systems with renewables , in: booktitle Smart Grid Communications (SmartGridComm), 2011 IEEE International Conference on , organization IEEE . pp. pages 457--462

work page 2011

[8] [8]

, author Neal, R

author Farivar, M. , author Neal, R. , author Clarke, C. , author Low, S. , year 2012 . title Optimal inverter var control in distribution systems with high pv penetration , in: booktitle Power and Energy Society General Meeting, 2012 IEEE , organization IEEE . pp. pages 1--7

work page 2012