Sharif Digital Repository

We propose an ℓ1 criterion for dictionary learning for sparse signal representation. Instead of directly searching for the dictionary vectors, our dictionary learning approach identifies vectors that are orthogonal to the subspaces in which the training data concentrate. We study conditions on the coefficients of training data that guarantee that ideal normal vectors deduced from the dictionary are local optima of the criterion. We illustrate the behavior of the criterion on a 2D example, showing that the local minima correspond to ideal normal vectors when the number of training data is sufficient. We conclude by describing an algorithm that can be used to optimize the criterion in higher... 

In this paper, based on a successively accuracy-increasing approximation of the ℓ0 norm, we propose a new algorithm for recovery of sparse vectors from underdetermined measurements. The approximations are realized with a certain class of concave functions that aggressively induce sparsity and their closeness to the ℓ0 norm can be controlled. We prove that the series of the approximations asymptotically coincides with the ℓ1 and ℓ0 norms when the approximation accuracy changes from the worst fitting to the best fitting. When measurements are noise-free, an optimization scheme is proposed that leads to a number of weighted ℓ1 minimization programs, whereas, in the presence of noise, we propose... 

In this paper, we develop and analyze a novel performance metric, called interference efficiency (IE), that shows the number of transmitted bits per unit of interference energy imposed on the primary users (PUs) in an underlay cognitive radio network (CRN). Specifically, we develop a framework to maximize the IE of a CRN with multiple secondary users (SUs) while satisfying target constraints on the average interference power on PU receiver, total SUs transmit power and minimum ergodic rate for the SUs. In doing so, we formulate a multiobjective optimization problem (MOP), that aims to achieve the maximum ergodic sum rate of multiple SUs and the minimum average interference power on the... 

The equilibrium signal-setting problem is stated and subsequently formulated as a continuous equilibrium network design problem. The bilevel formulation is nonconvex and therefore cannot be solved for global optima by using descent solution algorithms. Therefore, a lower bound using a system optimal flow pattern is proposed that will be quite tight in both uncongested and highly congested network traffic situations. A solution algorithm based on the standard steepest-descent method is proposed for the lower-bound problem. Performance of the solution algorithm on a network problem is reported  

In this paper, the intra-cell link adaptation problem is formulated as a constrained nonconvex nondifferentiable optimization problem to maximize the average link throughput while guaranteeing the best possible coverage reliability. We proceed to solve the resulted nonsmooth optimization problem using proximal point bundle method known in nondifferentiable optimization context. For this purpose, we first exploit the problem structure and reduce the original large scale optimization problem to a sequence of one-dimensional problems coordinated by a master program using direct primal decomposition technique. Proximal bundle method with aggregation policy is then adopted in master program to... 

In this paper, the assignability of linear time-invariant (LTI) systems with respect to arbitrary control structures is addressed. It is well established that the closed-loop spectrum of an LTI system with an arbitrary control structure is confined to the set containing the fixed-modes of the system with respect to that control structure. However, the assignment of the closed-loop spectrum is not merely limited by the existence of fixed-modes in practical scenarios. The pole assignment may require excessive control effort or even become infeasible due to the presence of small perturbations in the system dynamics. To offer more insights in such more realistic scenarios, a continuous measure... 

Unmanned aerial vehicles (UAVs) are developing rapidly owing to flexible deployment and access services as the air base stations. In this paper, by taking the realistic assumption of imperfect channel state information at transmitter (CSIT), we investigate the robust and secure design of downlink UAV network while considering the worst-case outage constraints due to communication link uncertainties. In our proposed heterogeneous network, comprised of both UAV-cells and macro-users, rate splitting technique is deployed by UAV base-station (UAV-BS) to enhance the system performance in terms of security, power saving, and robustness against imperfect CSIT. Through our proposed design, we... 

In this paper, the assignability of linear time-invariant (LTI) systems with respect to arbitrary control structures is addressed. It is well established that the closed-loop spectrum of an LTI system with an arbitrary control structure is confined to the set containing the fixed-modes of the system with respect to that control structure. However, the assignment of the closed-loop spectrum is not merely limited by the existence of fixed-modes in practical scenarios. The pole assignment may require excessive control effort or even become infeasible due to the presence of small perturbations in the system dynamics. To offer more insights in such more realistic scenarios, a continuous measure... 

This paper presents a new robust control strategy for an islanded microgrid in the presence of load unmodeled dynamics. The microgrid consists of parallel connection of several electronically interfaced distributed generation units and a local load. The load is parametrically uncertain and topologically unknown and, thus, is the source of unmodeled dynamics. The objective is to design a robust controller to regulate the load voltage in the presence of unmodeled dynamics. To achieve the objective, the problem is first characterized by a two-degree-of-freedom (2DOF) feedback-feedforward controller. The 2DOF control design problem is then transformed to a nonconvex optimization problem.... 

Heat exchanger network synthesis (HENS) is one of the most efficient process integration tools to save energy in chemical plants. In this work, a new optimization framework is proposed for the synthesis of HENS, based on a genetic algorithm (GA) coupled with a commercial process simulator through the ActiveX capability of the simulator. The use of GA provides a robust search in complex and nonconvex spaces of mathematical problems, while the use of a simulator facilitates the formulation of rigorous models for different alternatives. To include the most common heat exchanger structures in the model, a promising superstructure has been used. Allowing nonisothermal mixing of streams in the new... 

In this paper, the resource allocation and scheduling problem for a full-duplex (FD) orthogonal frequency-division multiple-access network is studied where an FD base station simultaneously communicates with multiple pairs of uplink (UL) and downlink (DL) half-duplex (HD) users bidirectionally. In this paper, we aim to maximize the network sum-rate through joint UL and DL user pairing, OFDM subchannel assignment, and power allocation. We formulate the problem as a non-convex optimization problem. The optimal algorithm requires an exhaustive search, which will become prohibitively complicated as the numbers of users and subchannels increase. To tackle this complex problem more efficiently, we... 

In this paper we introduce a novel population-based binary optimization technique, which works based on consensus of interacting multi-agent systems. The agents, each associated with an opinion vector, are connected through a network. They can influence each other, and thus their opinions can be updated. The agents work collectively with their neighbors to solve an optimization task. Here we consider a specific opinion update rule and various topologies for the connection network. Our experiments on a number of benchmark non-convex cost functions show that ring topology results in the best performance as compared to others. We also compare the performance of the proposed method with a number... 

This letter presents a novel solution for the problem of moving target localization in multiple-input multiple-output radar systems. The localization problem is formulated, based on least squares criterion, as a non-convex optimization problem and solved by semidefinite relaxation method. Then, an improvement technique, refining the initial solution by estimating the error terms, is proposed. Numerical simulations demonstrate that the proposed method achieves a significant performance improvement over the state-of-the-art methods. Specifically, the proposed method is shown to be more robust to the noise level compared with the existing algorithms. © 1997-2012 IEEE  

This paper presents a novel heuristic algorithm for solving economic dispatch (ED) problems, by employing iteration particle swarm optimization with time varying acceleration coefficients (IPSO-TVAC) method. Due to the effect of valve-points and prohibited operation zones (POZs) in the generating units' cost functions, ED problem is a non-linear and non-convex optimization problem. The problem even may be more complicated if transmission losses are taken into account. The effectiveness of the proposed method is examined and validated by carrying out extensive tests on three different test systems. Valve-point effects, POZs, ramp-rate constraints and transmission losses are modeled. Numerical...