Achieving Geometric Convergence for Distributed Optimization Over Time-Varying Graphs
Open accessThe DOI matches its Crossref record; 89 of the 89 references it cites were found in a registry or an index; 10 of the 34 sources its claims cite were checked for support.
Abstract
This paper considers the problem of distributed optimization over time-varying graphs. For the case of undirected graphs, we introduce a distributed algorithm, referred to as DIGing, based on a combination of a distributed inexact gradient method and a gradient tracking technique. The DIGing algorithm uses doubly stochastic mixing matrices and employs fixed step-sizes and, yet, drives all the agents' iterates to a global and consensual minimizer. When the graphs are directed, in which case the implementation of doubly stochastic mixing matrices is unrealistic, we construct an algorithm that incorporates the push-sum protocol into the DIGing structure, thus obtaining the Push-DIGing algorithm. Push-DIGing uses column stochastic matrices and fixed step-sizes, but it still converges to a global and consensual minimizer. Under the strong convexity assumption, we prove that the algorithms converge at R-linear (geometric) rates as long as the step-sizes do not exceed some upper bounds. We establish explicit estimates for the convergence rates. When the graph is undirected it shows that DIGing scales polynomially in the number of agents. We also provide some numerical experiments to demonstrate the efficacy of the proposed algorithms and to validate our theoretical findings.
Overview
CoChat Summary
TL;DR
The paper introduces DIGing for distributed optimization over time-varying undirected graphs and Push-DIGing for time-varying directed graphs, both using fixed step-sizes and gradient tracking. Under strong convexity and smoothness, it proves global R-linear (geometric) convergence and gives explicit rate bounds.
Key findings
- DIGing uses doubly stochastic mixing matrices and fixed step-sizes; under Assumptions 1, 4, and 5, it converges to the unique optimum at a global R-linear rate \(O(\lambda^k)\) for step-sizes \(\alpha \in (0, 1.5(1-\delta)^2 \bar{\mu} J_1]\).
- Push-DIGing incorporates the push-sum protocol with column stochastic matrices and fixed step-sizes; under Assumptions 4–7, it also converges at a global R-linear rate \(O(\lambda^k)\) for \(\alpha \in (0, 1.5(1-\delta)^2 \bar{\mu} J_2]\).
- For undirected graphs, Corollary 11 gives iteration complexity \(O(\tau^{-2} B^3 n^{4.5} \bar{\kappa}^{1.5} \ln(1/\varepsilon))\), polynomial in the number of agents \(n\). For lazy Metropolis mixing on time-invariant connected undirected graphs, Corollary 12 gives rate \(\lambda = 1 - 1/(161312\, n^{4.5} \bar{\kappa}^{1.5})\) and iteration complexity \(O(n^{4.5} \bar{\kappa}^{1.5} \ln(1/\varepsilon))\).
- Numerical experiments with \(n=12\), \(m_i=1\), \(p=3\), and Huber loss show DIGing and variants have R-linear rates, while the push-gradient method has only sublinear rate; DIGing-ATC is faster in iterations than DIGing but has higher per-iteration communication cost, and Push-DIGing can be as fast as DIGing-ATC at half communication cost per iteration.
Methods
- The paper studies distributed convex optimization \(\min \sum_{i=1}^n f_i(x)\) over time-varying undirected and directed graphs, assuming smoothness with Lipschitz gradients and strong convexity of at least one local function.
- It proposes DIGing: \(x(k+1)=W(k)x(k)-\alpha y(k)\), \(y(k+1)=W(k)y(k)+\nabla f(x(k+1))-\nabla f(x(k))\), and Push-DIGing, which adds push-sum variables \(u,v,V\) and uses column stochastic matrices.
- Convergence analysis uses the small-gain theorem with an ergodic norm and a cycle of arrows \(q \to z \to \check{y} \to \check{x} \to q\); numerical experiments compare push-gradient, DIGing, DIGing-ATC, and Push-DIGing on time-invariant directed, time-varying undirected, and time-varying directed graphs.
Limitations
- The paper states an open question is obtaining improved estimates on convergence rates, especially scaling with the number of agents \(n\), and extensions to local constraints and couplings among agents.
- Step-size selection requires agents to agree on common upper bounds on \(n\) and \(B\), plus lower \(\bar{\mu}\) and upper \(L\); the paper assumes all agents use the same step-size, though it notes uncoordinated step-sizes are possible in subsequent work.
- The paper states its bounds on step-sizes are often conservative and uses hand-optimized step-sizes in experiments; the numerical experiments use a small network of \(n=12\), so the empirical validation is limited in scale.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
9 Valid · 0 Not valid · 10 Not checked · 37 Own results and interpretations
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“This compares favorably to some of the existing literature, e.g., [70,76], which require detailed spectral information about the network for step-size selection.”
What the sources say
- Supportedcites [70] C. Xi and U. Khan, On the Linear Convergence of Distributed Optimization over Directed Graphs , preprint,arXiv:1510.02149, 2015. source
The part of the sentence this citation is offered for — that the cited work's step-size selection needs detailed spectral information about the network — is borne out by our copy: Theorem 1 (p. 13) sets α_min and α_max in terms of λ_min(M+M^⊤), d_- and d_∞, and the discussion after it (p. 14) states outright that 'the values of α and ᾱ need the knowledge of the network topology, which may not be available in a distributed manner,' flagging avoidance of that global knowledge as an open question. Section VI (p. 22) demonstrates the requirement in practice, computing λ_max/λ_min of H_i^⊤H_i, the degree constants and λ_min(M+M^⊤) to estimate the interval [9.6×10^-4, 0.26] before picking α = 0.1. The comparative framing ('compares favorably') is the citing authors' own conclusion and is not attributed to this source, so it is not judged here. (LLM, conf=0.85)
- Partly supportedcites [76] J. Zeng and W. Yin, ExtraPush for Convex Smooth Decentralized Optimization over Directed Networks , preprint,arXiv:1511.02942, 2015. sourceExperimental — accuracy not yet measured
Theorem 4 (p. 8) gives a condition on the step size, and the discussion on p. 9 relates the convergence bound to eigenvalues of network-derived matrices. But p. 9 calls the step-size condition sufficient, not necessary, and §6.1 (p. 12) selects step sizes by hand. The paper therefore supports a spectral dependence in its theoretical bound, not the categorical claim that step-size selection requires detailed spectral information. (LLM, conf=0.79)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“Proposition 1 ( [62]). Assume null { I -W } = span { 1 } where W ∈ R n × n . If x ∗ satisfies conditions: (i) x ∗ = Wx ∗ (consensus), (ii) 1 ⊤ ∇ f ( x ∗ ) = 0 (optimality), then the rows of x ∗ are the same as each other and the transpose of each row is an optimal solution of the problem (1) .”
What the sources say
- Supportedcites [62] matched to “EXTRA: An Exact First-Order Algorithm for Decentralized Consensus Optimization” (2015) source
The paper’s discussion of Proposition 2.1 (p. 6) says that, under null{I−W} = span{1}, a point satisfying its consensus and optimality conditions has identical rows that solve problem (1.1). Lemma 3.1’s proof (p. 9) identifies 1ᵀ∇f(x*) = 0 as the optimality condition. The citing proposition paraphrases those conditions and conclusion. (LLM, conf=0.92)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“While the subgradient-push eliminates the requirement of graph balancing [22], it suffers from a slow sublinear 2 convergence rate even for strongly convex smooth functions due to its employment of diminishing step-size [41].”
What the sources say
- Supportedcites [41] A. Nedić and A. Olshevsky, Stochastic Gradient-Push for Strongly Convex Functions on Time-Varying Directed Graphs , preprint, arXiv:1406.2075, 2014. source
The paper reports an O((ln t)/t) rate for strongly convex functions with Lipschitz gradients using an appropriately chosen diminishing step-size (main-result passage; page unavailable). This is sublinear, although the Introduction and Conclusion emphasize that it improves on the earlier O((ln t)/√t) rate and is close to the best achievable rate in the noisy setting; “slow” is the citing authors’ characterization. (LLM, conf=0.85)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“Assumption 2 is typical for many results in multi-agent coordination and distributed optimization [42]. It is much weaker than the assumption of every G un ( k ) being connected.”
What the sources say
- Supportedcites [42] matched to “Distributed Optimization Over Time-Varying Directed Graphs” (2015) source
In II. Problem, Algorithm and Main Results (p. 2), the paper calls its connectivity assumption typical and says it is considerably weaker than requiring each graph to be connected, because the necessary edges can appear over time. (LLM, conf=0.91)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
A result the paper states — it cites no source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“When the graph is undirected it shows that DIGing scales polynomially in the number of agents.”
- Also stated in: § ACHIEVING GEOMETRIC CONVERGENCE FOR DISTRIBUTED OPTIMIZATION OVER TIME-VARYING GRAPHS ∗ · p. 4 · plain
“Moreover, we show that the convergence rate of DIGing scales polynomially in the total number of agents in the network.”
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Rests on 2 assumptions not shown here.
- Also stated in: § ACHIEVING GEOMETRIC CONVERGENCE FOR DISTRIBUTED OPTIMIZATION OVER TIME-VARYING GRAPHS ∗ · p. 4 · plain
25 more claims
Create a free account to see all 56 claims and to verify the 2 references still open.
- It may state two separate claims as one. 13
- Probably not a substantial claim. 14
- The quote only partly states this claim. 1
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Algorithm 2: Push-DIGing · p. 21
“N out i ( k ) is the set of agents that can receive the information from agent i (in-neighbors of agent i ) at time k .”
§ Algorithm 2: Push-DIGing · p. 21
“N out i ( k ) = { j ∣ ∣ ( - → i, j ) ∈ A ( k ) }”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Algorithm 1: DIGing · p. 10
“the gradients are close to zero in the neighborhood of the optimal point”
§ ACHIEVING GEOMETRIC CONVERGENCE FOR DISTRIBUTED OPTIMIZATION OVER TIME-VARYING GRAPHS ∗ · p. 5
“not necessarily a point where ∇ f i ( x ) = 0 ∀ i”
Citations
Cites
Boyd, S., Parikh, N., Chu, E., Peleato, B., & Eckstein, J. (2011) · Foundations and Trends® in Machine Learning · doi:10.1561/2200000016 · Cited by 17,075
Boyd, S. (2010) · now publishers, Inc. eBooks · doi:10.1561/9781601984616 · Cited by 12,677
Distributed Subgradient Methods for Multi-Agent Optimization
Nedic, A., & Ozdaglar, A. (2009) · IEEE Transactions on Automatic Control · doi:10.1109/tac.2008.2009515 · Cited by 3,719 · cited 2×
“References [60–62] use a difference structure to cancel the steady state error in decentralized gradient descent [47, 75], thereby developing the algorithm EXTRA and its proximal-gradient variant.”
MethodFast linear iterations for distributed averaging
Xiao, L., & Boyd, S. (2004) · Systems & Control Letters · doi:10.1016/j.sysconle.2004.02.022 · Cited by 2,706
Markov Chains and Mixing Times
Levin, D., Peres, Y., & Wilmer, E. (2008) · doi:10.1090/mbk/058 · Cited by 2,241
Distributed asynchronous deterministic and stochastic gradient optimization algorithms
Tsitsiklis, J., Bertsekas, D., & Athans, M. (1986) · IEEE Transactions on Automatic Control · doi:10.1109/tac.1986.1104412 · Cited by 1,967
Gossip-based computation of aggregate information
Kempe, D., Dobra, A., & Gehrke, J. (2004) · 44th Annual IEEE Symposium on Foundations of Computer Science, 2003. Proceedings. · doi:10.1109/sfcs.2003.1238221 · Cited by 1,480 · cited 4×
“Instead, if every agent knows its out-degree, it is possible for the agents to construct a column stochastic matrix C and perform the following recursions with initialization u(0) = x(0) and v(0) = 1 to achieve the average (push-sum protocol [27]):”
BackgroundMethodInfluentialEXTRA: An Exact First-Order Algorithm for Decentralized Consensus Optimization
Shi, W., Ling, Q., Wu, G., & Yin, W. (2015) · SIAM Journal on Optimization · doi:10.1137/14096668x · Cited by 1,261
Self-citationDistributed Optimization Over Time-Varying Directed Graphs
Nedic, A., & Olshevsky, A. (2015) · IEEE Transactions on Automatic Control · doi:10.1109/tac.2014.2364096 · Cited by 1,187
Self-citationDistributed average consensus with least-mean-square deviation
Xiao, L., Boyd, S., & Kim, S.-J. (2007) · Journal of Parallel and Distributed Computing · doi:10.1016/j.jpdc.2006.08.010 · Cited by 1,107 · cited 1×
“Some examples include distributed averaging [7,50,72], distributed spectrum sensing [1], formation control [49, 58], power system control [21, 55], statistical inference and learning [20, 45, 53].”
Background
Research works printed but not linked: 9
- Chen A., 2012, Allerton, 601
- Di Lorenzo P., 2015, Proceedings of the IEEE International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, 229
- Forero P., 2010, J. Mach. Learn. Res., 59, 1663
- Mokhtari A., 2016, J. Mach. Learn. Res., 17, 1
- Nedić A., 2008, UK, 3
- A. Olshevsky, Efficient Information Aggregation Strategies for Distributed Control and Signal Processing , Ph.D. thesis, Massachusetts Institute of Technology, 2010.
- Rabbat M., 2004, New York, 20
- Ram S., 2009, Proceedings of INFOCOM, 3001
- Sayed A., 2013, New York, 323
Cited by
Items 301–400 of 1,070
Distributed Online Private Learning of Convex Nondecomposable Objectives
Cheng, H., Liao, X., & Li, H. (2023) · IEEE Transactions on Network Science and Engineering · doi:10.1109/tnse.2023.3329832 · Cited by 18
“1 Related Works Distributed offline learning: The study of distributed offline algorithm is relatively mature, such as subgradientpush [4], primal-dual [5], [6], gradient tracking [7], [8], [9], subgradient rescaling [10], asynchronous optimization [11], [12], etc.”
BackgroundWang, D., Liu, J., Lian, J., & Wang, W. (2024) · Automatica · doi:10.1016/j.automatica.2024.111925 · Cited by 14
SUCAG: Stochastic Unbiased Curvature-aided Gradient Method for Distributed Optimization
Wai, H.-T., Freris, N. M., Nedić, A., & Scaglione, A. (2018) · IEEE Conference on Decision and Control (CDC) · doi:10.1109/cdc.2018.8619336 · Cited by 17
“Moreover, we compare the DIG-ing method [27], an average consensusbased distributed optimization method that accounts for time varying communication topologies.”
MethodSelf-citationA Distributed Dual Proximal Minimization Algorithm for Constraint-Coupled Optimization Problems
Falsone, A., & Prandini, M. (2020) · IEEE Control Systems Letters · doi:10.1109/lcsys.2020.3001427 · Cited by 15
“Linear convergence rate for distributed algorithms with constant step-size solving DCPs has been recently achieved by the so-called gradient-tracking schemes (see [11], [12]), where the gradient method is used together with a technique known as dynamic average consensus (firstly proposed in [13] and further elaborated…”
MethodA Distributed Stochastic Proximal-Gradient Algorithm for Composite Optimization
Niu, Y., Li, H., Wang, Z., Lü, Q., Xia, D., & Ji, L. (2021) · IEEE Transactions on Control of Network Systems · doi:10.1109/tcns.2021.3065653 · Cited by 17
“…α 1 = α 2 = 1 and utilizing the saddle-point method with true gradient, we can recover the primal-dual interpretation of DIGing algorithm given by [18] as follows: Remark 3: In smooth cases, we can build the connections between the proposed algorithm and EXTRA, DSA, and DIG-ing algorithms by…”
BackgroundMethodInfluentialNetwork Gradient Descent Algorithm for Decentralized Federated Learning
Wu, S., Huang, D., & Wang, H. (2022) · Journal of Business and Economic Statistics · doi:10.1080/07350015.2022.2074426 · Cited by 22
“Nedic, Olshevsky, and Shi (2017) and Savazzi, Nicoli, and Rampa (2020) further proposed new algorithms based on the exchange of both parameters and gradients in each iteration for better convergence speed.”
BackgroundImproving the Transient Times for Distributed Stochastic Gradient Methods
Huang, K., & Pu, S. (2022) · IEEE Transactions on Automatic Control · doi:10.1109/tac.2022.3201141 · Cited by 28
“Problem (1.1) has been studied extensively in the literature using various distributed algorithms [32, 18, 19, 13, 6, 7, 28, 5, 26, 17, 34, 25], among which the distributed gradient descent (DGD) method proposed in [18] has drawn the greatest attention.”
InfluentialAsynchronous Decentralized Optimization in Directed NetworksPreprint
Zhang, J., & You, K. (2019) · arXiv (Cornell University) · Cited by 17
A Non-Asymptotic Analysis of Network Independence for Distributed Stochastic Gradient DescentPreprint
Olshevsky, A., Paschalidis, I. C., & Pu, S. (2019) · arXiv (Cornell University) · Cited by 13
Self-citationProjected Gradient Method for Decentralized Optimization over Time-Varying NetworksPreprint
Rogozin, A., & Gasnikov, A. V. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1911.08527 · Cited by 21
“Research Faculty Engagement Program. papers with theoretical analysis of rate of convergence for convex type problems: Nedić and Olshevsky (2014); Nedić et al. (2017); Maros and Jaldén (2018); Lü et al. (2018); Van Scoy and Lessard (2019).”
BackgroundLi, H., Lin, Z., & Fang, Y. (2020) · arXiv (Cornell University) · Cited by 14
Distributed Adaptive Gradient Algorithm With Gradient Tracking for Stochastic Nonconvex Optimization
Han, D., Liu, K., Lin, Y., & Xia, Y. (2024) · IEEE Transactions on Automatic Control · doi:10.1109/tac.2024.3380710 · Cited by 16
“Moreover, the weighted adjacency matrix A is doubly stochastic and satisfies ρ ( A Remark 2: Assumption 3 can be satisfied in connected undirected networks through properly designed weight setting protocols, such as the Metropolis weight protocol [43].”
BackgroundPrivacy-Preserving Dynamic Average Consensus via Random Number Perturbation
Gao, L., Zhou, Y., Chen, X., et al. (2022) · IEEE Transactions on Circuits & Systems II Express Briefs · doi:10.1109/tcsii.2022.3219929 · Cited by 14
Hayashi, N., Sugiura, T., Kajiyama, Y., & Takai, S. (2019) · IET Control Theory and Applications · doi:10.1049/iet-cta.2019.0377 · Cited by 16
“The extensions for directed and time-varying communication networks have also been considered in a large number of research articles [7–9].”
BackgroundNested Distributed Gradient Methods with Adaptive Quantized Communication
Berahas, A. S., Iakovidou, C., & Wei, E. (2019) · IEEE Conference on Decision and Control (CDC) · doi:10.1109/cdc40024.2019.9029541 · Cited by 15
A General Framework of Exact Primal-Dual First-Order Algorithms for Distributed Optimization
Mansoori, F., & Wei, E. (2019) · IEEE Conference on Decision and Control (CDC) · doi:10.1109/cdc40024.2019.9029902 · Cited by 14
“Although different variations of the above iteration have linear convergence guarantees [23], [31], the linear convergence of MM is shown to be faster [19].”
BackgroundMethodA decentralized algorithm for large scale min-max problems
Mukherjee, S., & Chakraborty, M. (2020) · IEEE Conference on Decision and Control (CDC) · doi:10.1109/cdc42340.2020.9304470 · Cited by 16
Li, B., Li, Z., & Chi, Y. (2022) · SIAM Journal on Mathematics of Data Science · doi:10.1137/21m1450677 · Cited by 26
“2: initialization: set x Set the new parameter estimate x ( t ) = u ( t − 1) ,S .”
A Linear Algorithm for Quantized Event-Triggered Optimization Over Directed Networks
Yang, Y., Shi, L., & He, W. (2022) · IEEE/CAA Journal of Automatica Sinica · doi:10.1109/jas.2022.105614 · Cited by 14
Distributed hybrid optimization for multi-agent systems
Tan, X., Yang, Y., He, W., Cao, J., & Huang, T. (2022) · Science China Technological Sciences · doi:10.1007/s11431-022-2060-7 · Cited by 15
“After that, this method was extended to deal with the distributed optimization problem over time-varying graphs [29].”
MethodComputational Convergence Analysis of Distributed Optimization Algorithms for Directed Graphs
Zhang, S., Yi, X., George, J., & Yang, T. (2019) · International Conference on Control and Automation · doi:10.1109/icca.2019.8899565 · Cited by 14
“The first class of algorithms [11], [12], [26]–[29] are based on the push-sum method [30]–[32], which relaxes the requirement of doubly stochastic mixing matrices in distributed optimization algorithms for undirected graphs to column stochastic matrices.”
MethodDistributed fixed point method for solving systems of linear algebraic equations
Jakovetić, D., Krejić, N., Jerinkić, N. K., et al. (2021) · Automatica · doi:10.1016/j.automatica.2021.109924 · Cited by 14
“It is therefore of interest to compare the approach of solving (1) applying some distributed optimization method [11,16,18] to the minimization of the quadratic function 12x tAx−btx with DFIX.”
MethodOptimal Hydrogen Production Dispatch of Networked Hydrogen-Based Microgrids via a Distributed Method
He, W., Yu, J., Cao, C., et al. (2025) · IEEE Transactions on Sustainable Energy · doi:10.1109/tste.2025.3541097 · Cited by 13
“Inspired by consensus of marginal hydrogen production ben-efit of each electrolyzer, it is expected to apply the DIGing algorithm in [34] to solve the HPDP.”
A Distributed Nesterov-Like Gradient Tracking Algorithm for Composite Constrained Optimization
Zheng, L., Li, H., Li, J. Q., Wang, Z., Lü, Q., Shi, Y., Wang, H., Dong, T., Ji, L., & Xia, D. (2023) · IEEE Transactions on Signal and Information Processing over Networks · doi:10.1109/tsipn.2023.3239698 · Cited by 17
“More remarkable works can be found in [19], [20], [21], [22], [23].”
BackgroundDistributed cooperative reinforcement learning for multi‐agent system with collision avoidance
Lan, X., Yan, J., He, S., Zhao, Z., & Zou, T. (2023) · International Journal of Robust and Nonlinear Control · doi:10.1002/rnc.6985 · Cited by 13
Optimized Gradient Tracking for Decentralized Online Learning
Sharma, S. D., & Rajawat, K. (2024) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2024.3366437 · Cited by 17
“To mitigate these errors, various algorithms proposed in the literature utilize two classes of strategies: (a) combining consecutive DGD updates resulting in correction term(s), as in EXTRA [11] and NIDS [12], and (b) locally tracking the global gradient, as in DIGing and ATC-GT [13], [14].”
MethodDistributed Optimization With Asynchronous Computation and Event-Triggered Communication
Dong, Z., Jin, Y., Mao, S., Ren, W., Du, W., & Tang, Y. S. (2024) · IEEE Transactions on Automatic Control · doi:10.1109/tac.2024.3449140 · Cited by 18
“Specifically, there exists a Lipschitz constant L ∈ (0 , + ∞ ) such that the following inequality holds for any x, y ∈ R p : Assumption 2 (See [37]): For ∀ i ∈ V , the local cost function f i ( x ) is μ -strongly convex.”
On the Performance of Exact Diffusion over Adaptive Networks
Yuan, K., Alghunaim, S. A., Ying, B., & Sayed, A. H. (2019) · IEEE Conference on Decision and Control (CDC) · doi:10.1109/cdc40024.2019.9029537 · Cited by 10
“In addition to exact diffusion, there exist other bias-correction methods such as EXTRA [12], ESOM [13], NIDS [14], and gradient-tracking methods such as Aug-DGM [15], NEXT/SONATA [16], [17], DIGing [18], [19], and push-pull methods [20], [21].”
MethodBin, M., Notarnicola, I., & Parisini, T. (2022) · IEEE Conference on Decision and Control (CDC) · doi:10.1109/cdc51059.2022.9993284 · Cited by 14
Distributed Optimization with Gradient Descent and Quantized Communication
Rikos, A. I., Jiang, W., Charalambous, T., & Johansson, K. H. (2023) · IFAC-PapersOnLine · doi:10.1016/j.ifacol.2023.10.100 · Cited by 14
“In Nedic et al. (2017), the authors present the DIGing algorithm which achieves geometric convergence rate as long as some constraints over the fixed step size are fulfilled.”
BackgroundDecentralized Stochastic Gradient Tracking for Empirical Risk Minimization
Zhang, J., & You, K. (2019) · Cited by 15
Distributed Gradient Flow: Nonsmoothness, Nonconvexity, and Saddle Point Evasion
Swenson, B., Murray, R., Poor, H. V., & Kar, S. (2021) · IEEE Transactions on Automatic Control · doi:10.1109/tac.2021.3111853 · Cited by 14
“While a complete survey of this topic is beyond the scope of the paper, we note that key issues which have been addressed in this context include optimization over time-varying and directed communication networks [26], [27]; constrained optimization [28]–[30]; convergence rate analysis [31], [32]; and optimization of…”
BackgroundDistributed Stochastic Gradient Tracking MethodsPreprint
Pu, S., & Nedić, A. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1805.11454 · Cited by 13
Self-citationDistributed stochastic optimization with gradient tracking over strongly-connected networksPreprint
Xin, R., Sahu, A. K., Khan, U. A., et al. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1903.07266 · Cited by 13
“Methods for directed graphs that are based on gradient tracking [14]–[21] rely on separate iterations for eigenvector estimation that may impede the convergence.”
ASY-SONATA: Achieving Geometric Convergence for Distributed Asynchronous Optimization.
Tian, Y., Sun, Y., Du, B., & Scutari, G. (2018) · Cited by 14
Communication-efficient hierarchical distributed optimization for multi-agent policy evaluation
Ren, J., Haupt, J., & Guo, Z. (2020) · Journal of Computational Science · doi:10.1016/j.jocs.2020.101280 · Cited by 13
Convergence Analysis of Distributed Gradient Descent Algorithms With One and Two Momentum Terms
Liu, B., Chai, L., & Yi, J. (2022) · IEEE Transactions on Cybernetics · doi:10.1109/tcyb.2022.3218663 · Cited by 16
“Based on the gradient tracking method, Nedi´c et al. [48] constructed the DIGing algorithm for undirected graphs and the Push-DIGing algorithm for directed graphs.”
MethodWang, D., Zhou, J., Wen, G., Lü, J., & Chen, G. (2023) · IEEE Transactions on Network Science and Engineering · doi:10.1109/tnse.2023.3274559 · Cited by 13
“Assumption 3 is used to ensure the feasibility of the event-triggered controller design, which is a common assump-tionintheliterature[7], [10] as pointedout in[36].”
BackgroundMethodAn enhanced gradient-tracking bound for distributed online stochastic convex optimization
Alghunaim, S. A., & Yuan, K. (2023) · Signal Processing · doi:10.1016/j.sigpro.2023.109345 · Cited by 15
Accelerated Distributed Optimization Algorithm With Malicious Nodes
Zhang, S., Liu, Z., Wen, G., & Wang, Y. (2023) · IEEE Transactions on Network Science and Engineering · doi:10.1109/tnse.2023.3342165 · Cited by 15
“However, all the aforementioned distributed acceleration optimization algorithms [14], [15], [16] require the corresponding transition matrix to be symmetric and doubly stochastic.”
BackgroundMethodInfluentialDecentralized bilevel optimization
Chen, X., Huang, M., & Ma, S. (2024) · Optimization Letters · doi:10.1007/s11590-024-02101-4 · Cited by 43
“The following assumptions will be used, which are standard in bilevel optimization [5, 9, 12, 13] and decentralized optimization literature [28, 23, 17, 34].”
BackgroundMethodResultInfluentialXu, L., Yi, X., Deng, C., Shi, Y., Chai, T., & Yang, T. (2024) · IEEE Transactions on Cybernetics · doi:10.1109/tcyb.2024.3384924 · Cited by 23
“When integrating the quantizer (7) with the distributed gradient tracking algorithm in [14], [15], and [45], each agent is required to access the gradient information.”
BackgroundMethodDifferentially Private Distributed Optimization Over Diagraphs With Application to Image Deblurring
Yang, Z., He, W., & Yang, Y. (2024) · IEEE Transactions on Control of Network Systems · doi:10.1109/tcns.2024.3395851 · Cited by 14
“Itisworth noting that Assumption 1 guarantees a unique optimal solution to problem (1) and is also a standard assumption for achieving first-order algorithms possessing linear convergence rates [9], [12], [13].”
Distributed Control of Islanded DC Microgrids: A Passivity-Based Game Theoretical Approach
Fu, Z., Cenedese, C., Cucuzzella, M., et al. (2024) · IEEE Transactions on Control Systems Technology · doi:10.1109/tcst.2024.3405711 · Cited by 12
“Nedic et al. [23] and Liu et al. [24] propose discrete-time algorithms converging with a linear rate to the consensual min-imizer.”
BackgroundChen, Z., & Wang, Y. (2024) · IEEE Transactions on Automatic Control · doi:10.1109/tac.2024.3488504 · Cited by 17
Fenchel dual gradient methods for distributed convex optimization over time-varying networks
Wu, X., & Lu, J. (2017) · IEEE Conference on Decision and Control (CDC) · doi:10.1109/cdc.2017.8264079 · Cited by 14
“Foundation of China under grant 61603254, the Shanghai Pujiang Program under grant 16PJ1406400, and the Natural Science Foundation of Shanghai under grant 16ZR1422500. April 6, 2018 DRAFT 2 algorithm [14], and the Push-DIGing algorithm [14] can be implemented over time-varying networks and still provide convergence ra…”
BackgroundMethodInfluentialFederated Edge Network Utility Maximization for a Multi-Server System: Algorithm and Convergence
Karakoc, N., Scaglione, A., Martin Reisslein, B., & Wu, R. (2022) · IEEE/ACM Transactions on Networking · doi:10.1109/tnet.2022.3156530 · Cited by 15
“This is a known behavior for fixed step size [46].”
BackgroundLiu, C., Zhou, Z., Pei, J., Zhang, Y., & Shi, Y. (2022) · IEEE Transactions on Automatic Control · doi:10.1109/tac.2022.3209951 · Cited by 15
“DIGing [28], Harnessing [35], and AugDGM [25] are based on similar strategies, for which the convergence rate is Θ((1− β)(2)/κ).”
BackgroundHu, J., Chen, G., Li, H., Shen, Z., & Zhang, W. (2022) · IEEE Transactions on Network Science and Engineering · doi:10.1109/tnse.2022.3225229 · Cited by 13
“Especially when compared with the batch gradient algorithms EXTRA [16] and GT-DGD [17], [18], the accelerated performance is more obvious.”
InfluentialDistributed Random Reshuffling Over Networks
Huang, K., Li, X., Milzarek, A., Pu, S., & Qiu, J. (2023) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2023.3262181 · Cited by 16
Distributed Optimization Over Networks
Nedić, A. (2018) · Lecture notes in mathematics · doi:10.1007/978-3-319-97142-1_1 · Cited by 13
Self-citationDistributed Learning Over Networks: Effect of Using Historical Observations
Shi, C., & Yang, G. (2020) · IEEE Transactions on Automatic Control · doi:10.1109/tac.2020.2973583 · Cited by 13
“Assumption 1 is a common assumption to characterize a timevarying network, which has been adopted in the existing works, such as [1], [20], and [28].”
BackgroundCommunication-Efficient Frank-Wolfe Algorithm for Nonconvex Decentralized Distributed Learning
Xian, W., Huang, F., & Huang, H. (2021) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v35i12.17246 · Cited by 21
Zhang, M., Zhou, Y., Ge, Q., Zheng, R., & Wu, Q. (2021) · IEEE Transactions on Systems Man and Cybernetics Systems · doi:10.1109/tsmc.2021.3112691 · Cited by 12
Implicit Tracking-Based Distributed Constraint-Coupled Optimization
Li, J., & Su, H. (2022) · IEEE Transactions on Control of Network Systems · doi:10.1109/tcns.2022.3203486 · Cited by 17
“…to track the global gradient 1 n =1 ∇ h j ( x j ( t )) , which leads to the following distributed algorithm: It is worth mentioning that (7) is not a new algorithm essentially, since its discrete-time counterpart has been studied in [3] and [4], where the discrete-time DAC algorithm [20] is used.”
BackgroundChen, Z., & Li, N. (2022) · IEEE/CAA Journal of Automatica Sinica · doi:10.1109/jas.2022.105992 · Cited by 12
“Introduction D ISTRIBUTED optimization [1]–[4] has been widely applied to many scenarios such as smart grids [5], [6], traffic networks [7], [8], and sensor networks [9], [10].”
MethodFixed-time Distributed Optimization under Time-Varying Communication TopologyPreprint
Garg, K., Baranwal, M., Hero, A. O., & Panagou, D. (2019) · arXiv (Cornell University) · Cited by 11
IDEAL: Inexact DEcentralized Accelerated Augmented Lagrangian MethodPreprint
Arjevani, Y., Bruna, J., Can, B., Gürbüzbalaban, M., Jegelka, S., & Lin, H. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2006.06733 · Cited by 22
“Linearly convergent primal algorithms By and large, recent methods that achieve linear convergence in the strongly convex case [17, 25, 35, 36, 39, 46, 47, 50] can be shown to follow a general framework based on the augmented Lagrangian method, see Algorithm 1.”
BackgroundAchieving Linear Convergence in Federated Learning under Objective and Systems Heterogeneity.Preprint
Mitra, A., Jaafar, R. H., Pappas, G. J., & Hassani, H. (2021) · arXiv (Cornell University) · Cited by 14
Distributed Momentum-Based Multiagent Optimization With Different Constraint Sets
Zhou, X., Ma, Z., Zou, S., & Margellos, K. (2024) · IEEE Transactions on Automatic Control · doi:10.1109/tac.2024.3445575 · Cited by 12
“If we do not consider the momentum term, i.e. set β = 0 , Algorithm 1 can be roughly simplified to several existing methods that employ gradient averaging, such as these in [21], [28], [37].”
Chen, T.-Y., Chen, W., Hao, J., Yang, W., & Zhang, J. (2025) · IEEE Transactions on Evolutionary Computation · doi:10.1109/tevc.2025.3525713 · Cited by 15
“After that, DBO methods under different system characteristics have been extensively studied, including time-varying networks [9], unbalanced networks [10], continuous-time systems [11], constrained problems [10], etc. [12], [13].”
An empirical comparison of multi-agent optimization algorithms
Assran, M., & Rabbat, M. (2017) · IEEE Global Conference on Signal and Information Processing · doi:10.1109/globalsip.2017.8309024 · Cited by 11
On the Local Linear Rate of Consensus on the Stiefel Manifold
Chen, S., Garcia, A., Hong, M., & Shahrampour, S. (2023) · IEEE Transactions on Automatic Control · doi:10.1109/tac.2023.3330735 · Cited by 28
A Universal Decomposition for Distributed Optimization Algorithms
Scoy, B. V., & Lessard, L. (2022) · IEEE Control Systems Letters · doi:10.1109/lcsys.2022.3182640 · Cited by 15
“We first illustrate our results on the DIGing algorithm [3,18], which is described by the iterations”
MethodInfluentialOn the Convergence of Distributed Subgradient Methods under Quantization
Doan, T. T., Maguluri, S. T., & Romberg, J. (2018) · Allerton Conference on Communication, Control, and Computing · doi:10.1109/allerton.2018.8636036 · Cited by 14
“Distributed algorithms for these problems have received much attention in recent years, mostly focusing on three widely-studied algorithms, namely, the alternating direction method of multipliers (ADMM) [1], [11], [22], [23], distributed dual methods (mirror descent/dual averaging) [6], [7], [25], and distributed (sub…”
MethodA Unifying Approximate Method of Multipliers for Distributed Composite Optimization
Wu, X., & Lu, J. (2022) · IEEE Transactions on Automatic Control · doi:10.1109/tac.2022.3173171 · Cited by 21
“…AMM endowed with (38) still generalizes most existing algorithms discussed in Section III, including EXTRA [5], ID-FBBS [13], PGC [16], PG-EXTRA [12], DPGA [11], the decentralized ADMM in [4], D-FBBS [13], the distributed ADMM in [9], the distributed primal-dual algorithm in [14], and DIGing [26].”
BackgroundMethodResultInfluentialOptimization over time-varying directed graphs with row and column-stochastic matricesPreprint
Saadatniaki, F., Xin, R., & Khan, U. A. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1810.07393 · Cited by 11
“classification problem in a distributed manner, agents initialize their states xi 0 according to IID zero-mean Gaussian random variables with variance 9. The performance of TV-ABalong with Push-DIGing [26], and subgradientpush [18] (with constant and diminishing step-sizes) is shown in Fig. 3 with the average residual…”
BackgroundMethodInfluentialAn introduction to decentralized stochastic optimization with gradient trackingPreprint
Xin, R., Kar, S., & Khan, U. A. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1907.09648 · Cited by 11
“he push-sum (type) algorithms [34] that enable consensus with non-DS weights with the help of eigenvector estimation. Combining push-sum respectively with DSGD [11], [12], EXTRA [23], and GT-DGD [16]–[18] leads to SGP [35], DEXTRA [36], and ADD-OPT [37] that are applicable to arbitrary directed graphs. A similar idea…”
MethodDistributed Adaptive Newton Methods with Globally Superlinear ConvergencePreprint
Zhang, J., You, K., & Başar, T. (2020) · arXiv (Cornell University) · Cited by 11
Li, B., Cen, S., Chen, Y., & Chi, Y. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1909.05844 · Cited by 134
“…gradient tracking ideas [Zhu and Martínez, 2010] have been incorporated to adjust DGD to ensure its linear convergence using a constant step size [Nedić et al., 2017, Di Lorenzo and Scutari, 2016, Qu and Li, 2018]; see further developments in [Scutari and Sun, 2019, Sun et al., 2019, Xin et al.,…”
MethodDecentralized Accelerated Gradient Methods With Increasing Penalty ParametersPreprint
Li, H., Fang, C., Yin, W., et al. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1810.01053 · Cited by 11
“s step and then follows a subgradient descent with a diminishing step-size. To avoid the diminishing step-size, three different types of methods have been proposed. The first type of methods [8], [18]–[20] relies on tracking differences of gradients, which keeps a variable to estimate the average gradient and uses this…”
MethodFast decentralized non-convex finite-sum optimization with recursive variance reductionPreprint
Xin, R., Khan, U. A., & Kar, S. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2008.07428 · Cited by 11
“A dynamic fusion mechanism, called gradient tracking [10,15,23,32,51], removes ζ(2) by tracking the average of the local gradient estimators in GT-SARAH to learn the global gradient at each node.”
A Survey of Distributed Algorithms for Aggregative Games
Li, H., Li, J., Ran, L., Zheng, L., & Huang, T. (2025) · IEEE/CAA Journal of Automatica Sinica · doi:10.1109/jas.2024.124998 · Cited by 12
Zhao, Y., Liao, X., & He, X. (2023) · IEEE Transactions on Control of Network Systems · doi:10.1109/tcns.2023.3237496 · Cited by 10
“These issues can be solved by introducing the gradient tracking scheme [19], [20], [21].”
BackgroundNaderi, S., & Blondin, M. J. (2023) · IEEE Access · doi:10.1109/access.2023.3341295 · Cited by 14
“These extensions include the DIGing algorithm [59], EXact firsT-ordeR Algorithm (EXTRA) [60], and Network InDependent Step-size algorithm (NIDS) [61].”
BackgroundMethodHuang, Y., Meng, Z., & Sun, J. Q. (2024) · IEEE Transactions on Systems Man and Cybernetics Systems · doi:10.1109/tsmc.2024.3405453 · Cited by 10
“In [44], an equivalent relationship between the primal-dual framework algorithm and gradient tracking algorithms in [5] and [6] were established for the unconstrained optimization problems.”
BackgroundDistributed Stochastic Projection-Free Algorithm for Constrained Optimization
Jiang, X., Zeng, X., Xie, L., et al. (2024) · IEEE Transactions on Automatic Control · doi:10.1109/tac.2024.3481040 · Cited by 10
“Toobtain abetterconvergencewithnondiminishingstep-sizes,gradient trackingtechniqueisintegratedinthedesignofdistributed optimizationalgorithms[9],[10],[11].”
Consensus-Based Distributed Optimization for Multiagent Systems Over Multiplex Networks
Rodríguez-Camargo, C. D., Urquijo-Rodríguez, A. F., & Mojica‐Nava, E. (2024) · IEEE Transactions on Control of Network Systems · doi:10.1109/tcns.2024.3510602 · Cited by 15
“We introduce the augmented supra-Lagrangian concept, which incorporates implicit gradient tracking [23], [33] for both intralayer and interlayer gradients.”
Lü, Q., Li, H., Liao, X., & Li, H. (2017) · International Conference on Information Science and Technology · doi:10.1109/icist.2017.7926783 · Cited by 10
Distributed Optimization over General Directed Networks with Random Sleep Scheme
Wang, Z., Zheng, L., & Li, H. (2020) · International Journal of Control Automation and Systems · doi:10.1007/s12555-018-9543-9 · Cited by 10
Payoff Distribution in Robust Coalitional Games on Time-Varying Networks
Raja, A. A., & Grammatico, S. (2021) · IEEE Transactions on Control of Network Systems · doi:10.1109/tcns.2021.3113269 · Cited by 10
The Barzilai–Borwein Method for distributed optimization over unbalanced directed networks
Hu, J., Chen, X., Zheng, L., Zhang, L., & Li, H. (2021) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2020.104151 · Cited by 13
“The restriction on step-size is relaxed by the follow-up works ADD-OPT/Push-DIGing [26, 19], AB [28], and AB m [29], where AB simultaneously utilizes both row-and column-stochastic weight matrices to gain an accelerated convergence rate.”
BackgroundMethodInfluentialNear-Optimal Decentralized Algorithms for Saddle Point Problems over Time-Varying Networks
Beznosikov, A. N., Rogozin, A., Kovalev, D. A., & Gasnikov, A. V. (2021) · Lecture notes in computer science · doi:10.1007/978-3-030-91059-4_18 · Cited by 15
“In particular, in B-connected networks [24] the graph can be disconnected at some times, but the union of any B consequent graphs must be connected.”
BackgroundExponentially convergent distributed Nash equilibrium seeking for constrained aggregative games
Liang, S., Yi, P., Hong, Y., et al. (2022) · Autonomous Intelligent Systems · doi:10.1007/s43684-022-00024-4 · Cited by 10
“Although exponential convergence results have been obtained in some literature on distributed optimization and game such as [11, 17, 18], the exponential results cannot hold for projected dynamics that deal with local constraints.”
BackgroundProximal Stochastic Recursive Momentum Methods for Nonconvex Composite Decentralized Optimization
Mancino-Ball, G., Miao, S., Xu, Y., & Chen, J. (2023) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v37i7.26087 · Cited by 11
“They include methods that work for convex problems only, such as DGD (Yuan, Ling, and Yin 2016), EXTRA (Shi et al. 2015), ADMM (Shi et al. 2014), DIG-ing (Nedic, Olshevsky, and Shi 2017), Acc-DNGD (Qu and Li 2019), MSDA (Scaman et al. 2017), DPAG (Ye et al. 2020), Flex-PD (Mansoori and Wei 2021), IDEAL (Arje-vani et a…”
BackgroundMethodTian, Y., & How, J. P. (2023) · IEEE Transactions on Robotics · doi:10.1109/tro.2023.3327635 · Cited by 9
“In the broader field of optimization, related works include decentralized consensus optimization methods such as [40], [41], [42], [43].”
MethodWang, Y. (2023) · IEEE Conference on Decision and Control (CDC) · doi:10.1109/cdc49753.2023.10383541 · Cited by 13
“For example, in dynamic-average-consensus based distributed optimization, the reference signals of individual agents are gradients of individual objective functions, which gradually converge to constant values as the iterates converge to the optimal solution [44].”
BackgroundExact spectral-like gradient method for distributed optimization
Jakovetić, D., Krejić, N., & Jerinkić, N. K. (2019) · Computational Optimization and Applications · doi:10.1007/s10589-019-00131-8 · Cited by 11
Event-Triggered Consensus-Based Optimization Algorithm for Smooth and Strongly Convex Cost Functions
Hayashi, N., Sugiura, T., Kajiyama, Y., & Takai, S. (2018) · IEEE Conference on Decision and Control (CDC) · doi:10.1109/cdc.2018.8618863 · Cited by 12
Distributed multi-step subgradient optimization for multi-agent system
Li, C., Chen, S., Li, J., & Wang, F. (2019) · Systems & Control Letters · doi:10.1016/j.sysconle.2019.04.008 · Cited by 11
Wang, J., Li, H., & Wang, Z. (2019) · IET Generation Transmission & Distribution · doi:10.1049/iet-gtd.2019.0144 · Cited by 11
Decentralized Stochastic Non-Convex Optimization over Weakly Connected Time-Varying Digraphs
Lu, S., & Wu, C. W. (2020) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp40776.2020.9053238 · Cited by 20
“This type of uniform bounds is also present in prior work such as [1, 15, 19].”
BackgroundDistributed accelerated optimization algorithms: Insights from an ODE
Chen, R., Yang, T., & Chai, T. (2020) · Science China Technological Sciences · doi:10.1007/s11431-020-1596-8 · Cited by 10
Hu, J., Xia, D., Cheng, H., Feng, L., Ji, L., Guo, J., & Li, H. (2020) · Asian Journal of Control · doi:10.1002/asjc.2483 · Cited by 10
“Therefore, we use the notion of the sufficiently small (positive) constant step-size and parameter in this paper, which is common in the literature [18,39].”
MethodConvergence of an accelerated distributed optimisation algorithm over time‐varying directed networks
Hu, J., Yan, Y., Li, H., Wang, Z., Xia, D., & Guo, J. (2020) · IET Control Theory and Applications · doi:10.1049/cth2.12022 · Cited by 10
A Unified Algorithmic Framework for Distributed Composite Optimization
Xu, J., Tian, Y., Sun, Y., & Scutari, G. (2020) · IEEE Conference on Decision and Control (CDC) · doi:10.1109/cdc42340.2020.9303791 · Cited by 10
“All the schemes in [1]– [4], [7]–[9], [11], [12], [14] are special cases of Algorithm (4) with specific choices of A, B, C ∈ W G .”
BackgroundResultInfluentialS-NEAR-DGD: A Flexible Distributed Stochastic Gradient Method for Inexact Communication
Iakovidou, C., & Wei, E. (2022) · IEEE Transactions on Automatic Control · doi:10.1109/tac.2022.3151734 · Cited by 13
“1 is that the two gradient tracking (GT) methods, namely EXTRA and DIGing, diverge without error correction (variants Q.2 and Q.3).”
BackgroundMethodInfluentialHamedani, E. Y., & Aybat, N. S. (2017) · Allerton Conference on Communication, Control, and Computing · doi:10.1109/allerton.2017.8262781 · Cited by 11
“, [29], [30], [31], [32], [33] considered static communication networks G = (N , E) while [34], [35] studied the time-varying networks.”
BackgroundWalkman: A Communication-Efficient Random-Walk Algorithm for Decentralized OptimizationPreprint
Mao, X., Yuan, K., Hu, Y., et al. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1804.06568 · Cited by 10
Gradient-Free Nash Equilibrium Seeking in N-Cluster Games with Uncoordinated Constant Step-SizesPreprint
Pang, Y., & Hu, G. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2008.13088 · Cited by 10
“Variations of such method have been extensively studied in distributed optimization, such as [49]–[53].”
Method
Metadata
- Authors
- Nedić, A., Olshevsky, A., & Shi, W.
- Affiliations
- Boston University (US)
- Arizona State University (US)
- Venue
- SIAM Journal on Optimization
- Published
- 2017-01
- Volume
- 27
- Issue
- 4
- Pages
- 2597-2633
- ISSN
- 1052-6234 (ISSN-L)
- 1095-7189
- OpenAlex
- W2963649943
- Cited by
- 1,069
- References
- 38
- Citations by year
- 2026: 95 · 2025: 147 · 2024: 130 · 2023: 124 · 2022: 118 · 2021: 114 · 2020: 138 · 2019: 106 · 2018: 64 · 2017: 29
- FWCI
- 46.51
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- SIAM Journal on Optimization (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- U.S. Air Force (AF FA95501510394)
- Office of Naval Research (N00014-12-1-0998)
- National Science Foundation (CNS 15-44953)
- Topics
- Physical Sciences › Engineering › Control and Systems Engineering › Distributed Control Multi-Agent Systems
- Physical Sciences › Computer Science › Computational Theory and Mathematics › Optimization and Variational Analysis
- Physical Sciences › Computer Science › Artificial Intelligence › Stochastic Gradient Optimization Techniques
- Keywords
- distributed optimization, dynamic graphs, geometric convergence, R-linear convergence, gradient tracking, strong convexity, fixed step size, undirected graphs, directed graphs
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (qwen3.8-omni-flash and gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.