The Experts below are selected from a list of 48 Experts worldwide ranked by ideXlab platform
Gesualdo Scutari - One of the best experts on this subject based on the ideXlab platform.
-
Achieving Linear Convergence in Distributed ASynchronous Multiagent Optimization
IEEE Transactions on Automatic Control, 2020Co-Authors: Ye Tian, Ying Sun, Gesualdo ScutariAbstract:This article studies multiagent (convex and nonconvex ) optimization over static digraphs. We propose a general distributed aSynchronous algorithmic framework whereby 1) agents can update their local variables as well as communicate with their neighbors at any time, without any form of coordination; and 2) they can perform their local computations using (possibly) delayed, out-of-Sync Information from the other agents. Delays need not be known to the agent or obey any specific profile, and can also be time-varying (but bounded). The algorithm builds on a tracking mechanism that is robust against aSynchrony (in the above sense), whose goal is to estimate locally the average of agents’ gradients. When applied to strongly convex functions, we prove that it converges at an R-linear (geometric) rate as long as the step-size is sufficiently small. A sublinear convergence rate is proved, when nonconvex problems and/or diminishing, uncoordinated step-sizes are considered. To the best of our knowledge, this is the first distributed algorithm with provable geometric convergence rate in such a general aSynchronous setting. Preliminary numerical results demonstrate the efficacy of the proposed algorithm and validate our theoretical findings.
-
ASY-SONATA: Achieving Geometric Convergence for Distributed ASynchronous Optimization.
2018Co-Authors: Ye Tian, Ying Sun, Gesualdo ScutariAbstract:Can one obtain a geometrically convergent algorithm for distributed aSynchronous multi-agent optimization? This paper provides a positive answer to this open question. The proposed algorithm solves multi-agent (convex and nonconvex) optimization over static digraphs and it is aSynchronous, in the following sense: i) agents can update their local variables as well as communicate with their neighbors at any time, without any form of coordination; and ii) they can perform their local computations using (possibly) delayed, out-of-Sync Information from the other agents. Delays need not obey any specific profile, and can also be time-varying (but bounded). The algorithm builds on a tracking mechanism that is robust against aSynchrony (in the above sense), whose goal is to estimate locally the average of agents' gradients. When applied to strongly convex functions, we prove that it converges at an R-linear (geometric) rate as long as the step-size is sufficiently small. A sublinear convergence rate is proved, when nonconvex problems and/or diminishing, uncoordinated step-sizes are considered. Preliminary numerical results demonstrate the efficacy of the proposed algorithm and validate our theoretical findings.
-
Achieving Linear Convergence in Distributed ASynchronous Multi-agent Optimization
arXiv: Optimization and Control, 2018Co-Authors: Ye Tian, Ying Sun, Gesualdo ScutariAbstract:This papers studies multi-agent (convex and \emph{nonconvex}) optimization over static digraphs. We propose a general distributed \emph{aSynchronous} algorithmic framework whereby i) agents can update their local variables as well as communicate with their neighbors at any time, without any form of coordination; and ii) they can perform their local computations using (possibly) delayed, out-of-Sync Information from the other agents. Delays need not be known to the agent or obey any specific profile, and can also be time-varying (but bounded). The algorithm builds on a tracking mechanism that is robust against aSynchrony (in the above sense), whose goal is to estimate locally the average of agents' gradients. When applied to strongly convex functions, we prove that it converges at an R-linear (geometric) rate as long as the step-size is {sufficiently small}. A sublinear convergence rate is proved, when nonconvex problems and/or diminishing, {\it uncoordinated} step-sizes are considered. To the best of our knowledge, this is the first distributed algorithm with provable geometric convergence rate in such a general aSynchronous setting. Preliminary numerical results demonstrate the efficacy of the proposed algorithm and validate our theoretical findings.
-
Allerton - ASY-SONATA: Achieving Linear Convergence in Distributed ASynchronous Multiagent Optimization
2018 56th Annual Allerton Conference on Communication Control and Computing (Allerton), 2018Co-Authors: Ye Tian, Ying Sun, Gesualdo ScutariAbstract:This papers studies multi-agent (convex and nonconvex) optimization over static digraphs. We propose a general distributed aSynchronous algorithmic framework whereby i) agents can update their local variables as well as communicate with their neighbors at any time, without any form of coordination; and ii) they can perform their local computations using (possibly) delayed, out-of-Sync Information from their neighbors. Delays need not be known to the agents or obey any specific profile, and can also be time-varying (but bounded). The algorithm builds on a tracking mechanism that is robust against aSynchrony (in the above sense), whose goal is to estimate locally the sum of agents’ gradients. When applied to strongly convex functions, we prove that it converges at an R-linear (geometric) rate as long as the step-size is sufficiently small. A sublinear convergence rate is proved, when nonconvex problems and/or diminishing, uncoordinated step-sizes are employed. To the best of our knowledge, this is the first distributed algorithm with provable geometric convergence rate in such a general aSynchonous setting.
Ye Tian - One of the best experts on this subject based on the ideXlab platform.
-
Achieving Linear Convergence in Distributed ASynchronous Multiagent Optimization
IEEE Transactions on Automatic Control, 2020Co-Authors: Ye Tian, Ying Sun, Gesualdo ScutariAbstract:This article studies multiagent (convex and nonconvex ) optimization over static digraphs. We propose a general distributed aSynchronous algorithmic framework whereby 1) agents can update their local variables as well as communicate with their neighbors at any time, without any form of coordination; and 2) they can perform their local computations using (possibly) delayed, out-of-Sync Information from the other agents. Delays need not be known to the agent or obey any specific profile, and can also be time-varying (but bounded). The algorithm builds on a tracking mechanism that is robust against aSynchrony (in the above sense), whose goal is to estimate locally the average of agents’ gradients. When applied to strongly convex functions, we prove that it converges at an R-linear (geometric) rate as long as the step-size is sufficiently small. A sublinear convergence rate is proved, when nonconvex problems and/or diminishing, uncoordinated step-sizes are considered. To the best of our knowledge, this is the first distributed algorithm with provable geometric convergence rate in such a general aSynchronous setting. Preliminary numerical results demonstrate the efficacy of the proposed algorithm and validate our theoretical findings.
-
Achieving Linear Convergence in Distributed ASynchronous Multi-agent Optimization
2019Co-Authors: Ye Tian, Sun Ying, Scutari GesualdoAbstract:This papers studies multi-agent (convex and \emph{nonconvex}) optimization over static digraphs. We propose a general distributed \emph{aSynchronous} algorithmic framework whereby i) agents can update their local variables as well as communicate with their neighbors at any time, without any form of coordination; and ii) they can perform their local computations using (possibly) delayed, out-of-Sync Information from the other agents. Delays need not be known to the agent or obey any specific profile, and can also be time-varying (but bounded). The algorithm builds on a tracking mechanism that is robust against aSynchrony (in the above sense), whose goal is to estimate locally the average of agents' gradients. When applied to strongly convex functions, we prove that it converges at an R-linear (geometric) rate as long as the step-size is {sufficiently small}. A sublinear convergence rate is proved, when nonconvex problems and/or diminishing, {\it uncoordinated} step-sizes are considered. To the best of our knowledge, this is the first distributed algorithm with provable geometric convergence rate in such a general aSynchronous setting. Preliminary numerical results demonstrate the efficacy of the proposed algorithm and validate our theoretical findings.Comment: Part of this work has been presented to Allerton 2018; first version posted on arxiv on March 2018; revised Nov. 2018. To appear on IEEE Trans. on Automatic Contro
-
ASY-SONATA: Achieving Geometric Convergence for Distributed ASynchronous Optimization.
2018Co-Authors: Ye Tian, Ying Sun, Gesualdo ScutariAbstract:Can one obtain a geometrically convergent algorithm for distributed aSynchronous multi-agent optimization? This paper provides a positive answer to this open question. The proposed algorithm solves multi-agent (convex and nonconvex) optimization over static digraphs and it is aSynchronous, in the following sense: i) agents can update their local variables as well as communicate with their neighbors at any time, without any form of coordination; and ii) they can perform their local computations using (possibly) delayed, out-of-Sync Information from the other agents. Delays need not obey any specific profile, and can also be time-varying (but bounded). The algorithm builds on a tracking mechanism that is robust against aSynchrony (in the above sense), whose goal is to estimate locally the average of agents' gradients. When applied to strongly convex functions, we prove that it converges at an R-linear (geometric) rate as long as the step-size is sufficiently small. A sublinear convergence rate is proved, when nonconvex problems and/or diminishing, uncoordinated step-sizes are considered. Preliminary numerical results demonstrate the efficacy of the proposed algorithm and validate our theoretical findings.
-
Achieving Linear Convergence in Distributed ASynchronous Multi-agent Optimization
arXiv: Optimization and Control, 2018Co-Authors: Ye Tian, Ying Sun, Gesualdo ScutariAbstract:This papers studies multi-agent (convex and \emph{nonconvex}) optimization over static digraphs. We propose a general distributed \emph{aSynchronous} algorithmic framework whereby i) agents can update their local variables as well as communicate with their neighbors at any time, without any form of coordination; and ii) they can perform their local computations using (possibly) delayed, out-of-Sync Information from the other agents. Delays need not be known to the agent or obey any specific profile, and can also be time-varying (but bounded). The algorithm builds on a tracking mechanism that is robust against aSynchrony (in the above sense), whose goal is to estimate locally the average of agents' gradients. When applied to strongly convex functions, we prove that it converges at an R-linear (geometric) rate as long as the step-size is {sufficiently small}. A sublinear convergence rate is proved, when nonconvex problems and/or diminishing, {\it uncoordinated} step-sizes are considered. To the best of our knowledge, this is the first distributed algorithm with provable geometric convergence rate in such a general aSynchronous setting. Preliminary numerical results demonstrate the efficacy of the proposed algorithm and validate our theoretical findings.
-
Allerton - ASY-SONATA: Achieving Linear Convergence in Distributed ASynchronous Multiagent Optimization
2018 56th Annual Allerton Conference on Communication Control and Computing (Allerton), 2018Co-Authors: Ye Tian, Ying Sun, Gesualdo ScutariAbstract:This papers studies multi-agent (convex and nonconvex) optimization over static digraphs. We propose a general distributed aSynchronous algorithmic framework whereby i) agents can update their local variables as well as communicate with their neighbors at any time, without any form of coordination; and ii) they can perform their local computations using (possibly) delayed, out-of-Sync Information from their neighbors. Delays need not be known to the agents or obey any specific profile, and can also be time-varying (but bounded). The algorithm builds on a tracking mechanism that is robust against aSynchrony (in the above sense), whose goal is to estimate locally the sum of agents’ gradients. When applied to strongly convex functions, we prove that it converges at an R-linear (geometric) rate as long as the step-size is sufficiently small. A sublinear convergence rate is proved, when nonconvex problems and/or diminishing, uncoordinated step-sizes are employed. To the best of our knowledge, this is the first distributed algorithm with provable geometric convergence rate in such a general aSynchonous setting.
Ying Sun - One of the best experts on this subject based on the ideXlab platform.
-
Achieving Linear Convergence in Distributed ASynchronous Multiagent Optimization
IEEE Transactions on Automatic Control, 2020Co-Authors: Ye Tian, Ying Sun, Gesualdo ScutariAbstract:This article studies multiagent (convex and nonconvex ) optimization over static digraphs. We propose a general distributed aSynchronous algorithmic framework whereby 1) agents can update their local variables as well as communicate with their neighbors at any time, without any form of coordination; and 2) they can perform their local computations using (possibly) delayed, out-of-Sync Information from the other agents. Delays need not be known to the agent or obey any specific profile, and can also be time-varying (but bounded). The algorithm builds on a tracking mechanism that is robust against aSynchrony (in the above sense), whose goal is to estimate locally the average of agents’ gradients. When applied to strongly convex functions, we prove that it converges at an R-linear (geometric) rate as long as the step-size is sufficiently small. A sublinear convergence rate is proved, when nonconvex problems and/or diminishing, uncoordinated step-sizes are considered. To the best of our knowledge, this is the first distributed algorithm with provable geometric convergence rate in such a general aSynchronous setting. Preliminary numerical results demonstrate the efficacy of the proposed algorithm and validate our theoretical findings.
-
ASY-SONATA: Achieving Geometric Convergence for Distributed ASynchronous Optimization.
2018Co-Authors: Ye Tian, Ying Sun, Gesualdo ScutariAbstract:Can one obtain a geometrically convergent algorithm for distributed aSynchronous multi-agent optimization? This paper provides a positive answer to this open question. The proposed algorithm solves multi-agent (convex and nonconvex) optimization over static digraphs and it is aSynchronous, in the following sense: i) agents can update their local variables as well as communicate with their neighbors at any time, without any form of coordination; and ii) they can perform their local computations using (possibly) delayed, out-of-Sync Information from the other agents. Delays need not obey any specific profile, and can also be time-varying (but bounded). The algorithm builds on a tracking mechanism that is robust against aSynchrony (in the above sense), whose goal is to estimate locally the average of agents' gradients. When applied to strongly convex functions, we prove that it converges at an R-linear (geometric) rate as long as the step-size is sufficiently small. A sublinear convergence rate is proved, when nonconvex problems and/or diminishing, uncoordinated step-sizes are considered. Preliminary numerical results demonstrate the efficacy of the proposed algorithm and validate our theoretical findings.
-
Achieving Linear Convergence in Distributed ASynchronous Multi-agent Optimization
arXiv: Optimization and Control, 2018Co-Authors: Ye Tian, Ying Sun, Gesualdo ScutariAbstract:This papers studies multi-agent (convex and \emph{nonconvex}) optimization over static digraphs. We propose a general distributed \emph{aSynchronous} algorithmic framework whereby i) agents can update their local variables as well as communicate with their neighbors at any time, without any form of coordination; and ii) they can perform their local computations using (possibly) delayed, out-of-Sync Information from the other agents. Delays need not be known to the agent or obey any specific profile, and can also be time-varying (but bounded). The algorithm builds on a tracking mechanism that is robust against aSynchrony (in the above sense), whose goal is to estimate locally the average of agents' gradients. When applied to strongly convex functions, we prove that it converges at an R-linear (geometric) rate as long as the step-size is {sufficiently small}. A sublinear convergence rate is proved, when nonconvex problems and/or diminishing, {\it uncoordinated} step-sizes are considered. To the best of our knowledge, this is the first distributed algorithm with provable geometric convergence rate in such a general aSynchronous setting. Preliminary numerical results demonstrate the efficacy of the proposed algorithm and validate our theoretical findings.
-
Allerton - ASY-SONATA: Achieving Linear Convergence in Distributed ASynchronous Multiagent Optimization
2018 56th Annual Allerton Conference on Communication Control and Computing (Allerton), 2018Co-Authors: Ye Tian, Ying Sun, Gesualdo ScutariAbstract:This papers studies multi-agent (convex and nonconvex) optimization over static digraphs. We propose a general distributed aSynchronous algorithmic framework whereby i) agents can update their local variables as well as communicate with their neighbors at any time, without any form of coordination; and ii) they can perform their local computations using (possibly) delayed, out-of-Sync Information from their neighbors. Delays need not be known to the agents or obey any specific profile, and can also be time-varying (but bounded). The algorithm builds on a tracking mechanism that is robust against aSynchrony (in the above sense), whose goal is to estimate locally the sum of agents’ gradients. When applied to strongly convex functions, we prove that it converges at an R-linear (geometric) rate as long as the step-size is sufficiently small. A sublinear convergence rate is proved, when nonconvex problems and/or diminishing, uncoordinated step-sizes are employed. To the best of our knowledge, this is the first distributed algorithm with provable geometric convergence rate in such a general aSynchonous setting.
S. Chuprun - One of the best experts on this subject based on the ideXlab platform.
-
Preamble and embedded Synchronization for RF carrier frequency-hopped OFDM
IEEE Journal on Selected Areas in Communications, 2005Co-Authors: John Eric Kleider, G. Maalouli, S. Gifford, S. ChuprunAbstract:In this work, we apply RF carrier frequency hopping to orthogonal frequency-division multiplexing (FH-OFDM). Achievable hop rate and bandwidth efficiency are determined based on signal acquisition/Synchronization and data demodulation performance in the presence of unknown time-frequency offsets, and channel gain/phase perturbations. We compare performance using two different data-aided Synchronization approaches. The first method sends Synchronization Information in a preamble before the OFDM payload symbol, whereas the second method embeds the Synchronization Information directly into the OFDM symbol stream. In the embedded technique, superposition of the Synchronization Information causes interference onto the OFDM data Information. Thus, the Sync Information must be removed before satisfactory bit-error rate (BER) performance can be achieved. Consequently, embedded interference cancellation (EIC) is utilized which requires accurate estimation of the Synchronization offsets and channel perturbations. Using coherent quadrature phase-shift keying-OFDM modulation, performance comparisons are presented using the COST207 multipath fading channel model. Fading channel BER performance results indicate that the embedded technique incurs only a slight signal-to-noise ratio penalty (less than 1 dB) compared with the preamble method. However, the embedded method offers the potential for improved hop rate and bandwidth efficiency because no dedicated slot is required for a Synchronization field.
-
Channel Estimation Performance for Zero-Overhead Channel Access in Mobile Sensor Networks
2004Co-Authors: John Eric Kleider, S. Gifford, S. Chuprun, Brian SadlerAbstract:Abstract : In this work we study the effects of channel estimation/ equalization on the BER performance of RF carrier frequency hopped OFDM (FH-OFDM) when using Synchronization Information that is embedded directly into the OFDM baseband symbol stream. In sensor networks using CSMA, the acquisition preambles can require a large overhead percentage of the overall message. Embedding Synchronization Information into the data Information stream eliminates this channel access overhead, thus providing potential for zero-time overhead channel access. Superposition (embedding) of the Sync Information, however, causes interference onto the data Information, which must be removed for satisfactory BER performance. Consequently, embedded Synchronization interference cancellation (EIC) is utilized, which requires accurate channel state estimation. Using coherent 4QAM- and 16QAM-OFDM modulation, channel estimation and BER performance is evaluated using the COST207 multipath fading channel model. Less than a 1 dB performance difference is found between a preamble and embedded system for short message bursts (< 1msec) and a burst data rate of greater than 1.6 Mbit/sec. The channel estimation mean square error (MSE) versus pilot symbol overhead is also determined as a function of urban and rural channel environments.
-
Channel estimation performance for frequency hopped OFDM using embedded Synchronization
IEEE MILCOM 2004. Military Communications Conference 2004., 1Co-Authors: John Eric Kleider, G. Maalouli, S. Gifford, S. ChuprunAbstract:In this work we study the effects of channel estimation/equalization on the BER performance of RF carrier frequency hopped OFDM (FH-OFDM) when using Synchronization Information that is embedded directly into the OFDM time-domain baseband sample stream. Embedded Synchronization eliminates the need for dedicated preamble slots for signal acquisition/Synchronization, thus providing potential for improved channel access efficiency. Assuming perfect Synchronization of the unknown time-frequency offsets, demodulated bit error rate (BER) performance is evaluated in frequency selective fading channels. Superposition (embedding) of the Sync Information causes interference onto the OFDM data Information, which must be removed for satisfactory BER performance. Consequently, embedded Synchronization interference cancellation (EIC) is utilized, which requires accurate estimation of the channel state Information. Using coherent 4QAM-OFDM and 16QAM-OFDM modulation, BER performance is evaluated using the COST207 multipath fading channel model with RF carrier frequency hopping.
John Eric Kleider - One of the best experts on this subject based on the ideXlab platform.
-
Preamble and embedded Synchronization for RF carrier frequency-hopped OFDM
IEEE Journal on Selected Areas in Communications, 2005Co-Authors: John Eric Kleider, G. Maalouli, S. Gifford, S. ChuprunAbstract:In this work, we apply RF carrier frequency hopping to orthogonal frequency-division multiplexing (FH-OFDM). Achievable hop rate and bandwidth efficiency are determined based on signal acquisition/Synchronization and data demodulation performance in the presence of unknown time-frequency offsets, and channel gain/phase perturbations. We compare performance using two different data-aided Synchronization approaches. The first method sends Synchronization Information in a preamble before the OFDM payload symbol, whereas the second method embeds the Synchronization Information directly into the OFDM symbol stream. In the embedded technique, superposition of the Synchronization Information causes interference onto the OFDM data Information. Thus, the Sync Information must be removed before satisfactory bit-error rate (BER) performance can be achieved. Consequently, embedded interference cancellation (EIC) is utilized which requires accurate estimation of the Synchronization offsets and channel perturbations. Using coherent quadrature phase-shift keying-OFDM modulation, performance comparisons are presented using the COST207 multipath fading channel model. Fading channel BER performance results indicate that the embedded technique incurs only a slight signal-to-noise ratio penalty (less than 1 dB) compared with the preamble method. However, the embedded method offers the potential for improved hop rate and bandwidth efficiency because no dedicated slot is required for a Synchronization field.
-
Channel Estimation Performance for Zero-Overhead Channel Access in Mobile Sensor Networks
2004Co-Authors: John Eric Kleider, S. Gifford, S. Chuprun, Brian SadlerAbstract:Abstract : In this work we study the effects of channel estimation/ equalization on the BER performance of RF carrier frequency hopped OFDM (FH-OFDM) when using Synchronization Information that is embedded directly into the OFDM baseband symbol stream. In sensor networks using CSMA, the acquisition preambles can require a large overhead percentage of the overall message. Embedding Synchronization Information into the data Information stream eliminates this channel access overhead, thus providing potential for zero-time overhead channel access. Superposition (embedding) of the Sync Information, however, causes interference onto the data Information, which must be removed for satisfactory BER performance. Consequently, embedded Synchronization interference cancellation (EIC) is utilized, which requires accurate channel state estimation. Using coherent 4QAM- and 16QAM-OFDM modulation, channel estimation and BER performance is evaluated using the COST207 multipath fading channel model. Less than a 1 dB performance difference is found between a preamble and embedded system for short message bursts (< 1msec) and a burst data rate of greater than 1.6 Mbit/sec. The channel estimation mean square error (MSE) versus pilot symbol overhead is also determined as a function of urban and rural channel environments.
-
Channel estimation performance for frequency hopped OFDM using embedded Synchronization
IEEE MILCOM 2004. Military Communications Conference 2004., 1Co-Authors: John Eric Kleider, G. Maalouli, S. Gifford, S. ChuprunAbstract:In this work we study the effects of channel estimation/equalization on the BER performance of RF carrier frequency hopped OFDM (FH-OFDM) when using Synchronization Information that is embedded directly into the OFDM time-domain baseband sample stream. Embedded Synchronization eliminates the need for dedicated preamble slots for signal acquisition/Synchronization, thus providing potential for improved channel access efficiency. Assuming perfect Synchronization of the unknown time-frequency offsets, demodulated bit error rate (BER) performance is evaluated in frequency selective fading channels. Superposition (embedding) of the Sync Information causes interference onto the OFDM data Information, which must be removed for satisfactory BER performance. Consequently, embedded Synchronization interference cancellation (EIC) is utilized, which requires accurate estimation of the channel state Information. Using coherent 4QAM-OFDM and 16QAM-OFDM modulation, BER performance is evaluated using the COST207 multipath fading channel model with RF carrier frequency hopping.
