{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,9,11]],"date-time":"2025-09-11T19:30:05Z","timestamp":1757619005619,"version":"3.44.0"},"reference-count":33,"publisher":"Springer Science and Business Media LLC","issue":"1","license":[{"start":{"date-parts":[[2025,7,16]],"date-time":"2025-07-16T00:00:00Z","timestamp":1752624000000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2025,7,16]],"date-time":"2025-07-16T00:00:00Z","timestamp":1752624000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"DOI":"10.13039\/501100001862","name":"Svenska Forskningsr\u00e5det Formas","doi-asserted-by":"publisher","award":["2021-00306","2021-00306","2021-00306"],"award-info":[{"award-number":["2021-00306","2021-00306","2021-00306"]}],"id":[{"id":"10.13039\/501100001862","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100004359","name":"Vetenskapsr\u00e5det","doi-asserted-by":"publisher","award":["2021-04772","2021-04772","2021-04772"],"award-info":[{"award-number":["2021-04772","2021-04772","2021-04772"]}],"id":[{"id":"10.13039\/501100004359","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100004270","name":"Royal Institute of Technology","doi-asserted-by":"crossref","id":[{"id":"10.13039\/501100004270","id-type":"DOI","asserted-by":"crossref"}]}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["EURASIP J. Adv. Signal Process."],"abstract":"<jats:title>Abstract<\/jats:title>\n          <jats:p>For its well-established convergence properties, simplicity, and applicability to various optimization problems, the alternating direction method of multipliers (ADMM) has been at the center of several research fields. When applied to problems such as consensus optimization in distributed systems, ADMM is often implemented in a centralized manner, which may be challenging due to its dependency on the location and capacity of the node. While decentralized implementations have been proposed, these schemes require all worker nodes to either replicate the synchronization work of all nodes or execute their tasks in sequence, making them computationally and communication-wise expensive or slow. To tackle these challenges, we introduce two novel decentralized ADMM algorithms to enable decentralized learning without incurring redundant computations, excessive message transmissions, or extended convergence time. Through theoretical analysis, we have shown that our algorithms inherit the well-established convergence properties of the classical centralized ADMM while keeping the computational and communication complexity for each node at <jats:inline-formula>\n              <jats:alternatives>\n                <jats:tex-math>$${\\mathcal {O}}(1)$$<\/jats:tex-math>\n                <mml:math xmlns:mml=\"http:\/\/www.w3.org\/1998\/Math\/MathML\">\n                  <mml:mrow>\n                    <mml:mi>O<\/mml:mi>\n                    <mml:mo>(<\/mml:mo>\n                    <mml:mn>1<\/mml:mn>\n                    <mml:mo>)<\/mml:mo>\n                  <\/mml:mrow>\n                <\/mml:math>\n              <\/jats:alternatives>\n            <\/jats:inline-formula>. Moreover, we have identified specific conditions under which each of our algorithms is guaranteed to converge faster than the other as well as the classical centralized ADMM. Across several numerical simulations with various settings, our algorithms have converged several times faster than several state-of-the-art decentralized ADMM implementations.<\/jats:p>","DOI":"10.1186\/s13634-025-01225-8","type":"journal-article","created":{"date-parts":[[2025,7,16]],"date-time":"2025-07-16T08:08:41Z","timestamp":1752653321000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Fast-converging decentralized alternating direction method of multipliers for consensus optimization"],"prefix":"10.1186","volume":"2025","author":[{"given":"Jeannie","family":"He","sequence":"first","affiliation":[]},{"given":"Ming","family":"Xiao","sequence":"additional","affiliation":[]},{"given":"Mikael","family":"Skoglund","sequence":"additional","affiliation":[]}],"member":"297","published-online":{"date-parts":[[2025,7,16]]},"reference":[{"issue":"9","key":"1225_CR1","doi-asserted-by":"publisher","first-page":"1284","DOI":"10.3390\/e24091284","volume":"24","author":"M Xiao","year":"2022","unstructured":"M. Xiao, M. Skoglund, Coding for large-scale distributed machine learning. Entropy 24(9), 1284 (2022). https:\/\/doi.org\/10.3390\/e24091284","journal-title":"Entropy"},{"issue":"8","key":"1225_CR2","doi-asserted-by":"publisher","first-page":"2145","DOI":"10.1109\/TPDS.2017.2658620","volume":"28","author":"C Feng","year":"2017","unstructured":"C. Feng, H. Xu, B. Li, An alternating direction method approach to cloud traffic management. IEEE Trans. Parall. Distrib. Syst. 28(8), 2145\u20132158 (2017). https:\/\/doi.org\/10.1109\/TPDS.2017.2658620","journal-title":"IEEE Trans. Parall. Distrib. Syst."},{"issue":"1","key":"1225_CR3","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1561\/2200000016","volume":"3","author":"S Boyd","year":"2011","unstructured":"S. Boyd, N. Parikh, E. Chu, B. Peleato, J. Eckstein et al., Distributed optimization and statistical learning via the alternating direction method of multipliers. Foundations and Trends 3(1), 1\u2013122 (2011). https:\/\/doi.org\/10.1561\/2200000016","journal-title":"Foundations and Trends"},{"issue":"1","key":"1225_CR4","doi-asserted-by":"publisher","first-page":"108","DOI":"10.1186\/s13634-022-00942-8","volume":"2022","author":"L Fodor","year":"2022","unstructured":"L. Fodor, D. Jakoveti\u0107, D. Boberi\u0107 Krsti\u0107ev, S. \u0160krbi\u0107, A parallel ADMM-based convex clustering method. EURASIP J. Adv. Signal Process. 2022(1), 108 (2022). https:\/\/doi.org\/10.1186\/s13634-022-00942-8","journal-title":"EURASIP J. Adv. Signal Process."},{"issue":"6","key":"1225_CR5","doi-asserted-by":"publisher","first-page":"1289","DOI":"10.1109\/TPDS.2021.3110104","volume":"33","author":"F Zhang","year":"2022","unstructured":"F. Zhang, E. Xue, R. Guo, G. Qu, G. Zhao, A.Y. Zomaya, DS-ADMM++: A novel distributed quantized ADMM to speed up differentially private matrix factorization. IEEE Trans. Parallel Distrib. Syst. 33(6), 1289\u20131302 (2022). https:\/\/doi.org\/10.1109\/TPDS.2021.3110104","journal-title":"IEEE Trans. Parallel Distrib. Syst."},{"issue":"20","key":"1225_CR6","doi-asserted-by":"publisher","first-page":"5297","DOI":"10.1109\/TSP.2016.2593681","volume":"64","author":"K Huang","year":"2016","unstructured":"K. Huang, N.D. Sidiropoulos, Consensus-ADMM for general quadratically constrained quadratic programming. IEEE Trans. Signal Process. 64(20), 5297\u20135310 (2016). https:\/\/doi.org\/10.1109\/TSP.2016.2593681","journal-title":"IEEE Trans. Signal Process."},{"issue":"9","key":"1225_CR7","doi-asserted-by":"publisher","first-page":"4540","DOI":"10.1109\/TCYB.2019.2950337","volume":"51","author":"X Wang","year":"2019","unstructured":"X. Wang, J. Yan, B. Jin, W. Li, Distributed and parallel ADMM for structured nonconvex optimization problem. IEEE Trans. Cybern. 51(9), 4540\u20134552 (2019). https:\/\/doi.org\/10.1109\/TCYB.2019.2950337","journal-title":"IEEE Trans. Cybern."},{"key":"1225_CR8","unstructured":"Z. Xu, G. Taylor, H. Li, M.A. Figueiredo, X. Yuan, T. Goldstein, Adaptive consensus ADMM for distributed optimization. In Proceedings of the International Conference on Machine Learning(ICML), (2017) pp. 3841\u20133850"},{"key":"1225_CR9","doi-asserted-by":"publisher","DOI":"10.1109\/TII.2021.3126883","author":"P Singh","year":"2021","unstructured":"P. Singh, M. Masud, M.S. Hossain, A. Kaur, G. Muhammad, A. Ghoneim, Privacy-preserving serverless computing using federated learning for smart grids. IEEE Trans. Indust. Inform. (2021). https:\/\/doi.org\/10.1109\/TII.2021.3126883","journal-title":"IEEE Trans. Indust. Inform."},{"issue":"2","key":"1225_CR10","doi-asserted-by":"publisher","first-page":"339","DOI":"10.17775\/CSEEJPES.2019.01650","volume":"7","author":"M Wang","year":"2019","unstructured":"M. Wang, Y. Su, L. Chen, Z. Li, S. Mei, Distributed optimal power flow of DC microgrids: a penalty based ADMM approach. CSEE J. Power Energy Syst. 7(2), 339\u2013347 (2019). https:\/\/doi.org\/10.17775\/CSEEJPES.2019.01650","journal-title":"CSEE J. Power Energy Syst."},{"key":"1225_CR11","doi-asserted-by":"publisher","first-page":"786","DOI":"10.1109\/TSP.2023.3250839","volume":"71","author":"L Guo","year":"2023","unstructured":"L. Guo, X. Shi, J. Cao, Z. Wang, Decentralized inexact proximal gradient method with network-independent stepsizes for convex composite optimization. IEEE Trans. Signal Process. 71, 786\u2013801 (2023). https:\/\/doi.org\/10.1109\/TSP.2023.3250839","journal-title":"IEEE Trans. Signal Process."},{"key":"1225_CR12","doi-asserted-by":"publisher","first-page":"4282","DOI":"10.1109\/TSP.2021.3092347","volume":"69","author":"J Zhang","year":"2021","unstructured":"J. Zhang, H. Liu, A.M.-C. So, Q. Ling, A penalty alternating direction method of multipliers for convex composite optimization over decentralized networks. IEEE Trans. Signal Process. 69, 4282\u20134295 (2021). https:\/\/doi.org\/10.1109\/TSP.2021.3092347","journal-title":"IEEE Trans. Signal Process."},{"issue":"10","key":"1225_CR13","doi-asserted-by":"publisher","first-page":"2565","DOI":"10.1109\/TSP.2019.2907258","volume":"67","author":"Y Liu","year":"2019","unstructured":"Y. Liu, W. Xu, G. Wu, Z. Tian, Q. Ling, Communication-censored ADMM for decentralized consensus optimization. IEEE Trans. Signal Process. 67(10), 2565\u20132579 (2019). https:\/\/doi.org\/10.1109\/TSP.2019.2907258","journal-title":"IEEE Trans. Signal Process."},{"key":"1225_CR14","doi-asserted-by":"publisher","unstructured":"M. Ma, A.N. Nikolakopoulos, G.B. Giannakis, Fast decentralized learning via hybrid consensus ADMM. In Proceedings of the 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), (2018), pp. 3829\u20133833. https:\/\/doi.org\/10.1109\/ICASSP.2018.8462038","DOI":"10.1109\/ICASSP.2018.8462038"},{"issue":"22","key":"1225_CR15","doi-asserted-by":"publisher","first-page":"5930","DOI":"10.1109\/TSP.2016.2602803","volume":"64","author":"Q Ling","year":"2016","unstructured":"Q. Ling, Y. Liu, W. Shi, Z. Tian, Weighted ADMM for fast decentralized network optimization. IEEE Trans. Signal Process. 64(22), 5930\u20135942 (2016). https:\/\/doi.org\/10.1109\/TSP.2016.2602803","journal-title":"IEEE Trans. Signal Process."},{"issue":"10","key":"1225_CR16","doi-asserted-by":"publisher","first-page":"5170","DOI":"10.1109\/TCYB.2019.2953538","volume":"51","author":"B Wang","year":"2021","unstructured":"B. Wang, J. Fang, H. Duan, H. Li, Graph simplification-aided ADMM for decentralized composite optimization. IEEE Trans. Cybern. 51(10), 5170\u20135183 (2021). https:\/\/doi.org\/10.1109\/TCYB.2019.2953538","journal-title":"IEEE Trans. Cybern."},{"key":"1225_CR17","doi-asserted-by":"publisher","unstructured":"S. Shethia, A. Gupta, O. Thapliyal, I. Hwang, Distributed fast-tracking alternating direction method of multipliers (ADMM) algorithm with optimal convergence rate. In Proceedings of 2021 IEEE International Conference on Systems, Man, and Cybernetics (SMC), (2021). pp. 976\u2013981. https:\/\/doi.org\/10.1109\/SMC52423.2021.9658615","DOI":"10.1109\/SMC52423.2021.9658615"},{"key":"1225_CR18","doi-asserted-by":"publisher","first-page":"185","DOI":"10.1016\/j.ins.2019.08.039","volume":"507","author":"J Yan","year":"2020","unstructured":"J. Yan, F. Guo, C. Wen, G. Li, Parallel alternating direction method of multipliers. Inf. Sci. 507, 185\u2013196 (2020). https:\/\/doi.org\/10.1016\/j.ins.2019.08.039","journal-title":"Inf. Sci."},{"key":"1225_CR19","doi-asserted-by":"publisher","first-page":"300","DOI":"10.1016\/j.apm.2017.06.010","volume":"58","author":"G Chen","year":"2018","unstructured":"G. Chen, J. Li, A fully distributed ADMM-based dispatch approach for virtual power plant problems. Appl. Math. Model. 58, 300\u2013312 (2018). https:\/\/doi.org\/10.1016\/j.apm.2017.06.010","journal-title":"Appl. Math. Model."},{"issue":"2","key":"1225_CR20","doi-asserted-by":"publisher","first-page":"220","DOI":"10.1109\/TSIPN.2017.2681205","volume":"4","author":"M Maros","year":"2017","unstructured":"M. Maros, J. Jald\u00e9n, ADMM for distributed dynamic beamforming. IEEE Trans. Signal Inf. Process. Over Netw. 4(2), 220\u2013235 (2017). https:\/\/doi.org\/10.1109\/TSIPN.2017.2681205","journal-title":"IEEE Trans. Signal Inf. Process. Over Netw."},{"key":"1225_CR21","doi-asserted-by":"publisher","first-page":"18","DOI":"10.1109\/TSIPN.2019.2957719","volume":"6","author":"W Li","year":"2019","unstructured":"W. Li, Y. Liu, Z. Tian, Q. Ling, Communication-censored linearized ADMM for decentralized consensus optimization. IEEE Trans. Signal Inf. Process. Over Netw. 6, 18\u201334 (2019). https:\/\/doi.org\/10.1109\/TSIPN.2019.2957719","journal-title":"IEEE Trans. Signal Inf. Process. Over Netw."},{"key":"1225_CR22","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1186\/s13634-015-0213-2","volume":"2015","author":"Z Hu","year":"2015","unstructured":"Z. Hu, Y. Zhu, J. Xu, J. Wang, Y. Yang, Distributed weighted sum-rate maximization with multi-cell uplink-downlink throughput duality. EURASIP J. Adv. Signal Process. 2015, 1\u201316 (2015). https:\/\/doi.org\/10.1186\/s13634-015-0213-2","journal-title":"EURASIP J. Adv. Signal Process."},{"key":"1225_CR23","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1186\/1687-1499-2014-69","volume":"2014","author":"P Liu","year":"2014","unstructured":"P. Liu, J. Li, H. Li, K. Wang, Robust power allocation in two-tier heterogeneous networks. EURASIP J. Wirel. Commun. Netw. 2014, 1\u201313 (2014). https:\/\/doi.org\/10.1186\/1687-1499-2014-69","journal-title":"EURASIP J. Wirel. Commun. Netw."},{"key":"1225_CR24","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1186\/s13634-015-0226-x","volume":"2015","author":"T Erseghe","year":"2015","unstructured":"T. Erseghe, A distributed approach to the OPF problem. EURASIP J. Adv. Signal Process. 2015, 1\u201313 (2015). https:\/\/doi.org\/10.1186\/s13634-015-0226-x","journal-title":"EURASIP J. Adv. Signal Process."},{"key":"1225_CR25","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1186\/s13634-018-0589-x","volume":"2018","author":"M Ma","year":"2018","unstructured":"M. Ma, A.N. Nikolakopoulos, G.B. Giannakis, Hybrid ADMM: a unifying and fast approach to decentralized optimization. EURASIP J. Adv. Signal Process. 2018, 1\u201317 (2018). https:\/\/doi.org\/10.1186\/s13634-018-0589-x","journal-title":"EURASIP J. Adv. Signal Process."},{"key":"1225_CR26","doi-asserted-by":"publisher","first-page":"2513","DOI":"10.1109\/TSP.2020.2983167","volume":"68","author":"X Mao","year":"2020","unstructured":"X. Mao, K. Yuan, Y. Hu, Y. Gu, A.H. Sayed, W. Yin, Walkman: A communication-efficient random-walk algorithm for decentralized optimization. IEEE Trans. Signal Process. 68, 2513\u20132528 (2020). https:\/\/doi.org\/10.1109\/TSP.2020.2983167","journal-title":"IEEE Trans. Signal Process."},{"key":"1225_CR27","doi-asserted-by":"publisher","first-page":"5842","DOI":"10.1109\/TSP.2020.3027917","volume":"68","author":"Y Ye","year":"2020","unstructured":"Y. Ye, H. Chen, M. Xiao, M. Skoglund, H.V. Poor, Privacy-preserving incremental ADMM for decentralized consensus optimization. IEEE Trans. Signal Process. 68, 5842\u20135854 (2020). https:\/\/doi.org\/10.1109\/TSP.2020.3027917","journal-title":"IEEE Trans. Signal Process."},{"issue":"2","key":"1225_CR28","doi-asserted-by":"publisher","first-page":"391","DOI":"10.1109\/LCOMM.2019.2955442","volume":"24","author":"Y Ye","year":"2020","unstructured":"Y. Ye, H. Chen, Z. Ma, M. Xiao, Decentralized consensus optimization based on parallel random walk. IEEE Commun. Lett. 24(2), 391\u2013395 (2020). https:\/\/doi.org\/10.1109\/LCOMM.2019.2955442","journal-title":"IEEE Commun. Lett."},{"key":"1225_CR29","doi-asserted-by":"publisher","first-page":"69883","DOI":"10.1109\/ACCESS.2018.2880454","volume":"6","author":"F Wen","year":"2018","unstructured":"F. Wen, L. Chu, P. Liu, R.C. Qiu, A survey on nonconvex regularization-based sparse and low-rank recovery in signal processing, statistics, and machine learning. IEEE Access 6, 69883\u201369906 (2018). https:\/\/doi.org\/10.1109\/ACCESS.2018.2880454","journal-title":"IEEE Access"},{"key":"1225_CR30","doi-asserted-by":"publisher","unstructured":"W.J. Nash, T.L. Sellers, S.R. Talbot, A.J. Cawthorn, W.B. Ford, Abalone. UCI Machine Learning Repository (1995) https:\/\/doi.org\/10.24432\/C55C7W","DOI":"10.24432\/C55C7W"},{"key":"1225_CR31","doi-asserted-by":"publisher","unstructured":"Q. Li, Y. Diao, Q. Chen, B. He, Federated learning on non-iid data silos: An experimental study. In 2022 IEEE 38th International Conference on Data Engineering (ICDE), (2022), pp. 965\u2013978. https:\/\/doi.org\/10.1109\/ICDE53745.2022.00077","DOI":"10.1109\/ICDE53745.2022.00077"},{"issue":"1","key":"1225_CR32","doi-asserted-by":"publisher","first-page":"165","DOI":"10.1007\/s10107-016-1034-2","volume":"162","author":"M Hong","year":"2017","unstructured":"M. Hong, Z.-Q. Luo, On the linear convergence of the alternating direction method of multipliers. Math. Program. 162(1), 165\u2013199 (2017)","journal-title":"Math. Program."},{"key":"1225_CR33","doi-asserted-by":"publisher","unstructured":"J. He, M. Xiao, M. Skoglund, Fast-converging decentralized ADMM for consensus optimization. Proceedings of the IEEE International Conference on Artificial Intelligence in Engineering (CAI) (2024), pp. 575\u2013580. https:\/\/doi.org\/10.1109\/CAI59869.2024.00114","DOI":"10.1109\/CAI59869.2024.00114"}],"container-title":["EURASIP Journal on Advances in Signal Processing"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1186\/s13634-025-01225-8.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1186\/s13634-025-01225-8\/fulltext.html","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1186\/s13634-025-01225-8.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,9,7]],"date-time":"2025-09-07T11:41:13Z","timestamp":1757245273000},"score":1,"resource":{"primary":{"URL":"https:\/\/asp-eurasipjournals.springeropen.com\/articles\/10.1186\/s13634-025-01225-8"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,7,16]]},"references-count":33,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2025,12]]}},"alternative-id":["1225"],"URL":"https:\/\/doi.org\/10.1186\/s13634-025-01225-8","relation":{},"ISSN":["1687-6180"],"issn-type":[{"type":"electronic","value":"1687-6180"}],"subject":[],"published":{"date-parts":[[2025,7,16]]},"assertion":[{"value":"11 October 2024","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"6 May 2025","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"16 July 2025","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"Not applicable.","order":2,"name":"Ethics","group":{"name":"EthicsHeading","label":"Ethics approval and consent to participate"}},{"value":"The authors declare that they have no conflict of interest.","order":3,"name":"Ethics","group":{"name":"EthicsHeading","label":"Competing interests"}},{"value":"This is an extension of [33], a conference paper published in the Proceedings of the 2024 IEEE Conference on Artificial Intelligence as a conference paper. Compared to the conference paper version, this version has two algorithms instead of only one. In addition, this version has a significantly more extensive theoretical analysis as well as more thorough numerical experiments.","order":4,"name":"Ethics","group":{"name":"EthicsHeading","label":"Conference paper"}}],"article-number":"27"}}