Search

Musical onset detection is a key component in any beat tracking system. Existing onset detection methods are based on temporal/spectral analysis, or methods that integrate temporal and spectral information together with statistical estimation and machine learning models. In this paper, we propose a method to localize onset components in music by using the S-transform, and thus, the method is purely based on temporal/spectral data. Unlike the other methods based on temporal/spectral data, which usually rely on the short time Fourier transform (STET), our method enables effective isolation of crucial frequency subbands due to the frequency dependent resolution of S-transform. Moreover, numerical results show, even with less computationally intensive steps, the proposed method can closely resemble the performance of more resource intensive statistical estimation based approaches., QC 20190603

The need of fast distributed solvers for optimizationproblems in networked systems has motivated the recent developmentof the Fast-Lipschitz optimization framework. In such an optimization, problems satisfying certain qualifying conditions,such as monotonicity of the objective function and contractivityof the constraints, have a unique optimal solution obtained via fast distributed algorithms that compute the fixed point of the constraints. This paper extends the set of problems for which the Fast-Lipschitz framework applies. Existing assumptions on the problem form are relaxed and new and generalized qualifying conditions are established by novel results based on Lagrangianduality. It is shown for which cases of more constraints thandecision variables, and less constraints than decision variables Fast-Lipschitz optimization applies. New results are obtained by imposing non strict monotonicity of the objective functions. The extended Fast-Lipschitz framework is illustrated by a number ofexamples, including network optimization and optimal control problems., QC 20160111

In reactive cognitive networks, the channel access and the transmission decisions of the cognitive terminals have a long-term effect on the network dynamics. When multiple cognitive terminals coexist, the optimization and implementation of their strategy is challenging and may require considerable coordination overhead. In this paper, such challenge is addressed by a novel framework for the distributed optimization of transmission and channel access strategies. The objective of the cognitive terminals is to find the optimal action distribution depending on the current network state. To reduce the coordination overhead, in the proposed framework the cognitive terminals distributively coordinate the policy, whereas the action in each individual time slot is independently selected by the terminals. The optimization of the transmission and channel access strategy is performed iteratively by using the alternate convex optimization technique, where at each iteration a cognitive terminal is selected to optimize its own action distribution while assuming fixed those of the other cognitive terminals. For a traditional primary-secondary user network configuration, numerical results show that the proposed algorithm converges to a stable solution in a small number of iterations, and a limited performance loss with respect to the perfect coordinated case., QC 20161213

For operating electrical power networks, the Optimal Power Flow (OPF) problem plays a central role. The problem is nonconvex and NP hard. Therefore, designing efficient solution algorithms is crucial, though their global optimality is not guaranteed. Existing semi-definite programming relaxation based approaches are restricted to OPF problems for which zero duality holds, whereas for non-convex problems there is a lack of solution methods of provable performance. In this paper, an efficient novel method to address the general nonconvex OPF problem is investigated. The proposed method is based on alternating direction method of multipliers combined with sequential convex approximations. The global OPF problem is decomposed into smaller problems associated to each bus of the network, the solutions of which are coordinated via a light communication protocol. Therefore, the proposed method is highly scalable. The convergence properties of the proposed algorithm are mathematically and numerically substantiated., QC 20170201

The need of fast distributed solvers for optimization problems in networked systems has motivated the recent development of the Fast-Lipschitz optimization framework. In such an optimization, problems satisfying certain qualifying conditions, such as monotonicity of the objective function and contractivity of the constraints, have a unique optimal solution obtained via fast distributed algorithms that compute the fixed point of the constraints. This paper extends the set of problems for which the Fast-Lipschitz framework applies. Existing assumptions on the problem form are relaxed and new qualifying conditions are established. It is shown for which cases of more constraints than decision variables, and less constraints than decision variables Fast-Lipschitz optimization applies. New results are obtained by imposing non strict monotonicity of the objective functions. The extended Fast-Lipschitz framework is illustrated by a number of examples, including network optimization and optimal control problems, QC 20170201

Millimeter-wave communications in the 60-GHz band are considered one of the key technologies for enabling multigigabit wireless access. However, the special characteristics of such a band pose major obstacles to the optimal utilization of the wireless resources, where the problem of efficient client association to access points (APs) is of vital importance. In this paper, the client association in 60-GHz wireless access networks is investigated. The AP utilization and the quality of the rapidly vanishing communication links are the control parameters. Because of the tricky non-convex and combinatorial nature of the client association optimization problem, a novel solution method is developed to guarantee balanced and fair resource allocation. A new distributed, lightweight, and easy-to-implement association algorithm, based on Lagrangian duality theory and subgradient methods, is proposed. It is shown that the algorithm is asymptotically optimal, that is, the relative duality gap diminishes to zero as the number of clients increases., QC 20150703

This paper investigates convergence properties ofscalable algorithms for nonconvex and structured optimization.We consider a method that is adapted from the classic quadraticpenalty function method, the Alternating Direction PenaltyMethod (ADPM). Unlike the original quadratic penalty functionmethod, in which single-step optimizations are adopted, ADPMuses alternating optimization, which in turn is exploited toenable scalability of the algorithm. A special case of ADPM isa variant of the well known Alternating Direction Method ofMultipliers (ADMM), where the penalty parameter is increasedto infinity. We show that due to the increasing penalty, theADPM asymptotically reaches a primal feasible point undermild conditions. Moreover, we give numerical evidence thatdemonstrates the potential of the ADPM for computing localoptimal points when the penalty is not updated too aggressively., QC 20151202

Nonconvex and structured optimization problemsarise in many engineering applications that demand scalableand distributed solution methods. The study of the convergenceproperties of these methods is in general difficult due to thenonconvexity of the problem. In this paper, two distributedsolution methods that combine the fast convergence propertiesof augmented Lagrangian-based methods with the separabilityproperties of alternating optimization are investigated. The firstmethod is adapted from the classic quadratic penalty functionmethod and is called the Alternating Direction Penalty Method(ADPM). Unlike the original quadratic penalty function method,in which single-step optimizations are adopted, ADPM uses analternating optimization, which in turn makes it scalable. Thesecond method is the well-known Alternating Direction Methodof Multipliers (ADMM). It is shown that ADPM for nonconvexproblems asymptotically converges to a primal feasible pointunder mild conditions and an additional condition ensuringthat it asymptotically reaches the standard first order necessary conditions for local optimality are introduced. In thecase of the ADMM, novel sufficient conditions under whichthe algorithm asymptotically reaches the standard first ordernecessary conditions are established. Based on this, completeconvergence of ADMM for a class of low dimensional problemsare characterized. Finally, the results are illustrated by applyingADPM and ADMM to a nonconvex localization problem inwireless sensor networks., QC 20151208

The resource allocation problem of optimal assignment of the stations to the available access points in 60 GHz millimeterWave wireless access networks is investigated. The problem is posed as a multi-assignment optimization problem. The proposed solution method converts the initial problem to a minimum cost flow problem and allows to design an efficient algorithm by a combination of auction algorithms. The solution algorithm exploits the network optimization structure of the problem, and thus is much more powerful than computationally intensive general-purpose solvers. Theoretical and numerical results evince numerous properties, such as optimality, convergence, and scalability in comparison to existing approaches., QC 20151210

A framework for the distributed optimization of reactive cognitive networks with multiple secondary users is presented. The secondary users iteratively locate the policy maximizing their aggregate performance under bounds on the maximum performance loss caused to the primary users. The policy accounts for the impact of interference on the dynamics of the primary users' network due to transmission and channel access protocols. To minimize coordination overhead, it is assumed that the secondary users only coordinate the policy, whereas actions in each slot are independently selected by the individual secondary user based on the agreed policy. The distributed optimization technique proposed herein is based on alternating convex optimization. Numerical results are presented assessing the performance of the obtained transmission policy with respect to the optimal centralized and fully-coordinated policy., QC 20150612

The resource allocation problem of optimal assignment of the stations to the available access points in 60 GHz millimeterWave wireless access networks is investigated. The problem is posed as a multi-assignment optimization problem. The proposed solution method converts the initial problem to a minimum cost flow problem and allows to design an efficient algorithm by a combination of auction algorithms. The solution algorithm exploits the network optimization structure of the problem, and thus is much more powerful than computationally intensive general-purpose solvers. Theoretical and numerical results evince numerous properties, such as optimality, convergence, and scalability in comparison to existing approaches., QC 20140114

We develop an easy to implement distributed method for weighted sum-rate maximization (WSRMax) problem in a multicell multiple antenna downlink system. Unlike the recently proposed minimum weighted mean-squared error based algorithms, where at each iteration all mobile terminals needs to estimate the covariance matrices of their received signals, compute and feedback over the air certain parameters to the base stations (BS), our algorithm operates without any user terminal assistance. It requires only BS to BS signalling via reliable backhaul links (e.g. fiber, microwave links) and all required computation is performed at the BSs. The algorithm is based on primal decomposition and subgradient methods, where the original nonconvex problem is split into a master problem and a number of subproblems (one for each BS). A novel sequential convex approximation strategy is proposed to address the nonconvex master problem. In the case of subproblems, we adopt an existing iterative approach based on second-order cone programming and geometric programming. The subproblems are coordinated to find a (possibly suboptimal) solution to the master problem. Subproblems can be solved by BSs in a fully asynchronous manner, though the coordination between subproblems should be synchronous. Numerical results are provided to see the behavior of the algorithm under different degrees of BS coordination. They show that the proposed algorithm yields a good tradeoff between the implementation-level simplicity and the performance., QC 20140205

We develop an easy to implement distributed method for weighted sum-rate maximization (WSRMax) problem in a multicell multiple antenna downlink system. Unlike the recently proposed minimum weighted mean-squared error based algorithms, where at each iteration all mobile terminals needs to estimate the covariance matrices of their received signals, compute and feedback over the air certain parameters to the base stations (BS), our algorithm operates without any user terminal assistance. It requires only BS to BS signalling via reliable backhaul links (e.g., fiber, microwave links) and all required computation is performed at the BSs. The algorithm is based on primal decomposition and subgradient methods, where the original nonconvex problem is split into a master problem and a number of subproblems (one for each BS). A novel sequential convex approximation strategy is proposed to address the nonconvex master problem. In the case of subproblems, we adopt an existing iterative approach based on second-order cone programming and geometric programming. The subproblems are coordinated to find a (possibly suboptimal) solution to the master problem. Subproblems can be solved by BSs in a fully asynchronous manner, though the coordination between subproblems should be synchronous. Numerical results are provided to see the behavior of the algorithm under different degrees of BS coordination. They show that the proposed algorithm yields a good tradeoff between the implementation-level simplicity and the performance., QC 20130213

Wireless sensor networks for industrial automation applications must offer timely, reliable, and energy efficient communications at both low and high data rate. While traditional communication technologies between 2.4 GHz and 5 GHz are sometimes incapable to efficiently achieve the aforementioned goals, new communication strategies are emerging, such as millimeterWave communications. In this overview paper, the general requirements that factory and process automation impose on the network design are reviewed. Moreover, this paper presents and qualitatively evaluates the 60 GHz millimeterWave communication technology for automation. It is argued that the upcoming 60 GHz millimeterWave technology brings an enormous potential and can influence the design of the future communication infrastructures in factory and process automation., QC 20140110

Ensuring privacy is an essential requirement in various contexts, such as social networks, healthcare data, ecommerce, banks, and government services. Here, different entities coordinate to address specific problems where the sensitive problem data are distributed among the involved entities and no entity wants to publish its data during the solution procedure. Existing privacy preserving solution methods are mostly based on cryptographic procedures and thus have the drawback of substantial computational complexity. Surprisingly, little attention has been devoted thus far to exploit mathematical optimization techniques and their inherent properties for preserving privacy. Yet, optimization based approaches to privacy require much less computational effort compared to cryptographic variants, which is certainly desirable in practice. In this paper, a unified framework for transformation based optimization methods that ensure privacy is developed. A general definition for the privacy in the context of transformation methods is proposed. A number of examples are provided to illustrate the ideas. It is concluded that the theory is still in its infancy and that huge benefits can be achieved by a substantial development., QC 20140107

The problem of weighted sum-rate maximization (WSRMax) in multicell downlink multiple-input single-output (MISO) systems is considered. The problem is known to be NP-hard. We propose a method, based on branch and bound technique, which solves globally the nonconvex WSRMax problem with an optimality certificate. Specifically, the algorithm computes a sequence of asymptotically tight upper and lower bounds and it terminates when the difference between them falls below a pre-specified tolerance. Novel bounding techniques via conic optimization are introduced and their efficiency is demonstrated by numerical simulations. The proposed method can be used to provide performance benchmarks by back-substituting it into many existing network design problems which relies on WSRMax problem. The method proposed here can be easily extended to maximize any system performance metric that can be expressed as a Lipschitz continuous and increasing function of signal-to-interference-plus-noise ratio., QC 20120416

Ensuring privacy in distributed optimization is essential in many contexts, such as healthcare data,banks, e-commerce, government services, and social networks. In these contexts, it is common thatdifferent parties coordinate to solve a specific optimization problem whose data is dispersed amongthe parties, where no entity wants to publish its data during the solution procedure. Addressing theseproblems falls under the umbrella of well-knownsecured multiparty computation(SMC). Existingapproaches for SMC are mostly based on cryptography. Surprisingly, little attention has been devotedthus far to develop non-cryptographic approaches, that can be much more efficient. In this paper,we investigate alternative non-cryptographic methods based onmathematical optimization techniques.First, aunified frameworkto encapsulate existing non-cryptographic methods, which rely algebraictransformations to disguise sensitive problem data, is developed. The proposed framework capitalizes onkey optimization techniques, such aschange of variablesandtransformation of objective and constraintfunctions, for equivalent problem formation. Moreover, the privacy preserving properties that are inherentin the mathematical optimization techniques, including classical decomposition methods (e.g., primal anddual decomposition), and state-of-the-art methods, such as alternating direction method of multipliersare investigated. A general definition for quantifying the privacy in the context of non-cryptographicapproaches is proposed. A number of examples are provided to illustrate the importance of our proposedalgorithms. It is concluded that the theory is in its infancy and that huge benefits can be achieved by asubstantial development., QS 2014

QS 2016