Semi-asynchronous routing for large scale hierarchical networks

Liao, Wei-Cheng; Mei, Hong; Farmanbar, Hamid; Luo, Zhi-Quan

doi:10.1109/icassp.2015.7178500

Cited by 15 publications

(15 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, this task is difficult because the optimization variables of problem (32) is coupled (indeed each flow rate f m appears in exactly two flow conservation constraints). To address this problem, we introduce a few sets of new variables to decouple the flow conservation constraints across different subnetworks (we refer the readers to [95] for the detailed reformulation). The reformulated problem (32) is given by…”

Section: B Coupling Constraintsmentioning

confidence: 99%

See 1 more Smart Citation

A Unified Algorithmic Framework for Block-Structured Optimization Involving Big Data: With applications in machine learning and signal processing

Mei

Razaviyayn

Luo

et al. 2016

IEEE Signal Process. Mag.

Self Cite

416

266

View full text Add to dashboard Cite

show abstract

Section: B Coupling Constraintsmentioning

confidence: 99%

“…A wireline network consists of 5 subnetworks. Each of them is controlled by a NC, and these NCs are coordinated globally by a central NC 0[95].…”

mentioning

confidence: 99%

A Unified Algorithmic Framework for Block-Structured Optimization Involving Big Data: With applications in machine learning and signal processing

Mei

Razaviyayn

Luo

et al. 2016

IEEE Signal Process. Mag.

Self Cite

416

266

View full text Add to dashboard Cite

show abstract

“…Some recent work has derived an improvement for the distributed mechanism proposed in [8], which uses a standard gradient projection method for the decomposition approach, by employing alternating direction method of multipliers (ADMM) in order to obtain faster and more robust computation. See, e.g., [21]. It however requires substantial communication overhead, i.e., besides requiring each agent to know the cost and benefit of all possible choices, the agents need to know their share of unallocated capacity at each connected grid [22].…”

Section: B Related Workmentioning

confidence: 99%

“…As proposed in [17] and [31], we combine a dual decomposition approach with the gradient projection method to decouple problem (21). The feasibility and stability of the distributed MPC approach have been studied in [17] and [31] as well.…”

Section: B Distributed Mpc Problemmentioning

confidence: 99%

Asynchronous Distributed Control of Biogas Supply and Multienergy Demand

Alkano

Scherpen

Chorfi

2017

IEEE Trans. Automat. Sci. Eng.

View full text Add to dashboard Cite

In this paper we study the coordination between biogas producers who can either use the biogas themselves, exchange biogas with their neighbors, or deliver it to the various energy grids, such as the low pressure gas grid or the power grid. These producers are called prosumers. In this setting gas storage, fuel cells, micro combined heat power systems, and heat buffers are all part of the prosumers' node. We aim to optimize the imbalance, profit, and comfort levels per prosumer, while taking the constraints of the energy grids into account, and while allowing prosumers to exchange energy with each other. This results in a two-layer optimization problem formulation. In addition, in practice, communication between prosumers among each other and with grid operators is done in an asynchronous manner. In this paper we study the problem of two-layer optimization for biogas prosumers embedded in multiple energy grids, while the (bidirectional) communication between the various partners is done asynchronously. We prove the convergence of the asynchronous coordination algorithm that uses both the inputs and the states. We conduct simulations for the biogas prosumer setting, using realistic data to illustrate the convergence of the algorithm and to study its practical implementation. Note to Practitioners-This paper is motivated and supported by a smart gas grid project of the Energy Delta Gas Research (EDGaR) consortium in the Netherlands. The project deals with investigating the capacity of smart grid technologies to facilitate the introduction of new gases into the distribution grids, with diverse gas qualities and multiple injection points. The gas distribution grid will have to move from a passive to an active distribution system that dynamically control bidirectional flows between end-users and the grid operators. As the endusers may be equipped with energy converters, other energy distribution grids also need to transform to active distribution systems. Existing approaches are distributed, where each enduser and energy grid operator can locally solve their optimal control problem. In this paper, we consider the fact that both endusers and grid operators do not have access to a common clock when solving their problem and when sharing their information. The information includes some of their states and controllable inputs. The asynchronous information exchange problem was pointed out by DNV GL Netherlands, Gasunie, and Gasterra which are companies we collaborate with within the EDGaR consortium. It is highly relevant for practical implementation of our distributed algorithms. In future research, we will include practical control considerations due to on-off constraints of

show abstract

“…The rate here is w.r.t the minimization variable x; 2 Note that this algorithm has nested loops, so the total number of iterations are counted;3 The two results refer to the case where F is strongly concave and concave, respectively, w.r.t y.…”

mentioning

confidence: 99%

Hybrid Block Successive Approximation for One-Sided Non-Convex Min-Max Problems: Algorithms and Applications

Tsaknakis

Mei

et al. 2020

IEEE Trans. Signal Process.

Self Cite

View full text Add to dashboard Cite

The min-max problem, also known as the saddle point problem, is a class of optimization problems in which we minimize and maximize two subsets of variables simultaneously. This class of problems can be used to formulate a wide range of signal processing and communication (SPCOM) problems.Despite its popularity, existing theory for this class has been mainly developed for problems with certain special convex-concave structure. Therefore, it cannot be used to guide the algorithm design for many interesting problems in SPCOM, where some kind of non-convexity often arises.In this work, we consider a general block-wise one-sided non-convex min-max problem, in which the minimization problem consists of multiple blocks and is non-convex, while the maximization problem is (strongly) concave. We propose a class of simple algorithms named Hybrid Block Successive Approximation (HiBSA), which alternatingly performs gradient descent-type steps for the minimization blocks and one gradient ascent-type step for the maximization problem. A key element in the proposed algorithm is the introduction of certain properly designed regularization and penalty terms, which are used to stabilize the algorithm and ensure convergence. For the first time, we show that such simple alternating min-max algorithms converge to first-order stationary solutions, with quantifiable global rates. To validate the efficiency of the proposed algorithms, we conduct numerical tests on a number of information processing and wireless communication problems, including the robust learning problem, the non-convex min-utility maximization problems, and certain wireless jamming problem arising in interfering channels. * The first two authors contributed equally to this work, and the names are listed alphabetically. S. Lu, I. Tsaknakis and M. Hong are with the DRAFT Problem (1) is quite generic, and it arises in a wide range of signal processing and communication (SPCOM) applications. We list of few of these applications below. A. Motivating Examples in SPCOMDistributed non-convex optimization: Consider a network of K agents defined by a connected graph G = {V, E} with |V| = K, where each agent i can communicate with its neighbors. A generic problem formulation that captures many problems that appear in distributed machine learning and signal processing (e.g., [2]-[5]) can be formulated as follows:

show abstract

Semi-asynchronous routing for large scale hierarchical networks

Cited by 15 publications

References 12 publications

A Unified Algorithmic Framework for Block-Structured Optimization Involving Big Data: With applications in machine learning and signal processing

A Unified Algorithmic Framework for Block-Structured Optimization Involving Big Data: With applications in machine learning and signal processing

Asynchronous Distributed Control of Biogas Supply and Multienergy Demand

Hybrid Block Successive Approximation for One-Sided Non-Convex Min-Max Problems: Algorithms and Applications

Contact Info

Product

Resources

About