Adaptive algorithms play an important role in order to improve performance of diffusion distributed network. In comparison of diffusion Normalized least mean square algorithm, family of diffusion subband algorithms have faster convergence rate when the input signal is highly correlated. This paper solves the problem of distributed estimation in the diffusion networks based on improved multiband-structured subband adaptive filter (IMSAF) and diffusion improved multiband structured subband adaptive filter (DIMSAF) is established. In proposed algorithm, convergence behavior improved due to using several input projections instead of single vector of input data. In addition, when the projection order is increased, the convergence rate of the proposed algorithm improves. The validity of the DIMSAF in comparison of DLMS, DAPA, VSS-DAPA and DRLS algorithms is demonstrated by several computer simulations. The results show fastest convergence rate.

In this paper, the topology of complex adaptive networks based on distributed estimation least mean square (LMS) algorithm for the diffusion mode of cooperation (ATC) is studied. The study covers different network models, including the regular, small-world and random assuming temporal and spatial dependence of data. The parameters used for implementation of complex networks are average path length, cluster coefficient and algebraic connectivity in addition to the performance of the network. The simulation results indicate that in all complex networks with ideal and noisy links, the small-world networks (for adding random links) are better candidates for practical implementations due to high algebraic connectivity (robust to node failure problem), average path length and cluster coefficients (strong locality). Simulation results are also included for the Metropolis coefficient combination.

distributed processing uses local computations at each node and communications among neighboring nodes to solve the problems over the entire network. diffusion is one of the methods for performing distributed networks. This paper presents a novel Variable Step-Size diffusion Affine Projection Algorithm (VSS-DAPA) to improve the performance of the diffusion Affine Projection Algorithm (DAPA) in distributed networks. The variable step-size of each node is obtained by minimizing the Mean-Square Deviation (MSD) in that node. In comparison with diffusion Affine Projection Algorithm (DAPA), the VSS-DAPA algorithm has faster convergence speed and lower steady-state error. To reduce the computational complexity of VSS-DAPA, the Variable Step-Size Selective Regressors diffusion Affine Projection Algorithm (VSS-SR-DAPA), the Variable Step-Size Dynamic Selection of diffusion Affine Projection Algorithm (VSS-DS-DAPA) and Variable Step-Size Selective Partial Update diffusion Affine Projection Algorithm (VSS-SPU-DAPA) are proposed. Simulation results show the good performance of proposed algorithms in convergence speed and steady-state error.

Nowadays, most of information systems can be modeled as multilayer networks which each layer includes some nodes connected to each other by different types of links. Information diffusion in networks is the subject that researchers considered recently and they analyzed and modeled this process type in the networks. Although most of researches in this field have focused on single layer networks, but in the real world, because of the complexity of relations, most systems must be modeled as multilayer networks. In the previous works, there are much simplification in problem space, like projection all layer into one layer or negligence the mutual effect of nodes in different layers. So a new effective model for analyzing diffusion in multilayer networks is needed. This method is focused on predicting diffusion in multilayer networks, with considering mutual effect of different layers on each other. The most important specification of this proposed method, is the ability to specify power of all layers and measuring this power regardless node's similarity or difference. In fact, this model can determine the diffusion power of each type of nodes. The model is applied on two real bibliographic information networks, and experimentally demonstrated the effectiveness of this model compared with other diffusion models.

In this paper, a sensor network is used to estimate the dynamic states of a system. At each time step, one (or multiple) sensors are available that can send its measured data to a central node, in which all of processing is done. We want to provide an optimal algorithm for scheduling sensor selection at every time step. Our goal is to select the appropriate sensor to reduce computations, optimize the energy consumption and enhance the network lifetime. To achieve this goal, we must reduce the error covariance. Three algorithms are used in this work: sliding window, thresholding and randomly chosen algorithms. Moreover, we will offer a new algorithm based on circular selection. Finally, a novel algorithm for selecting multiple sensors is proposed. Performance of the proposed algorithms is illustrated with numerical examples.

In this paper, a novel distributed approach based on particle swarm optimization (PSO) has been proposed for distributed regression over sensor networks. Besides distributed data, the limitations of sensor nodes make doing regression more difficult. Conventional methods employ numerical optimization techniques such as gradient descent or Nelder-Mead Simplex algorithms in which the sensor nodes collaborate through a pre-established Hamiltonian path. Although NM Simplex based approaches converge faster than the gradient counterparts, both of them suffer from low accuracy and high latency. In the proposed approach, denoted as D-PSO (distributed PSO), a swarm of particles is dedicated to each cluster and the cluster regressor is learned. Afterwards, the clusters regressors are sent to the fusion center in order to build the global model. To do this, weighted averaging combination rule, which comes from MCS (Multiple Classifier Systems) concept, is applied on the received clusters regressors. The experimental results show that the proposed approach has a superior performance in terms of the prediction accuracy, latency, and energy efficiency compared to its counterparts.