Search

Time series prediction with deep learning methods, especially long short-term memory neural networks (LSTMs), have scored significant achievements in recent years. Despite the fact that the LSTMs can help to capture long-term dependencies, its ability to pay different degree of attention on sub-window feature within multiple time-steps is insufficient. To address this issue, an evolutionary attention-based LSTM training with competitive random search is proposed for multivariate time series prediction. By transferring shared parameters, an evolutionary attention learning approach is introduced to the LSTMs model. Thus, like that for biological evolution, the pattern for importance-based attention sampling can be confirmed during temporal relationship mining. To refrain from being trapped into partial optimization like traditional gradient-based methods, an evolutionary computation inspired competitive random search method is proposed, which can well configure the parameters in the attention layer. Experimental results have illustrated that the proposed model can achieve competetive prediction performance compared with other baseline methods.

Computer vision researchers have been expecting that neural networks have spatial transformation ability to eliminate the interference caused by geometric distortion for a long time. Emergence of spatial transformer network makes dream come true. Spatial transformer network and its variants can handle global displacement well, but lack the ability to deal with local spatial variance. Hence how to achieve a better manner of deformation in the neural network has become a pressing matter of the moment. To address this issue, we analyze the advantages and disadvantages of approximation theory and optical flow theory, then we combine them to propose a novel way to achieve image deformation and implement it with a hierarchical convolutional neural network. This new approach solves for a linear deformation along with an optical flow field to model image deformation. In the experiments of cluttered MNIST handwritten digits classification and image plane alignment, our method outperforms baseline methods by a large margin.

The problem of existence and stability of equilibria of DC microgirds with constant power loads methods is addressed in this paper. Constant power loads (CPLs) often cause instability due to its negative impedance characteristics. What is the worse, CPLs may cause that the system admits no equilibria for its nonlinearity. The purpose of paper can be summarized as: A) designing a controller to overcome CPLs instability; B) obtaining the sufficient conditions to guarantee the existence of the equilibria; C) Under the conditions of equilibria existing, obtaining the sufficient conditions of local stability. For the first objective, the method based on virtual resistance and virtual inductance is proposed. For the second objective, we transform the problem of solvability of quadratic equations into the existence of the fixed point of a mapping. Thus, the analytical sufficient conditions is obtained based on Tarski fixed point theorem. Moreover, a numerical method is also provided to reduce conservatism. For the third objective, the small-signal model is established to predict the system qualitative behavior around equilibria. The stability conditions are obtained by using quadratic eigenvalue problem (GEP) theorem. These explicit conditions are obtained as a function of system parameters and it lead to build reliable microgrid., Comment: 8 pages and 5 figures

This letter proposes a decentralized control scheme for grid-connected cascaded modular inverters without any communication, and each module makes decisions based on its own local information. In contrast, the conventional methods are usually centralized control and depend on a real-time communication. Thus, the proposed scheme has advantages of improved reliability and decreased costs. The overall system stability is analyzed, and the stability condition is derived as well. The feasibility of the proposed method is verified by simulation., Comment: 2 Pages, 2 figures

In order to address the economical dispatch problem in islanded microgrid, this letter proposes an optimal criterion and two decentralized economical-sharing schemes. The criterion is to judge whether global optimal economical-sharing can be realized via a decentralized manner. On the one hand, if the system cost functions meet this criterion, the corresponding decentralized droop method is proposed to achieve the global optimal dispatch. Otherwise, if the system does not meet this criterion, a modified method to achieve suboptimal dispatch is presented. The advantages of these methods are convenient,effective and communication-less., Comment: 3 pades, 3 figures

To address the economical dispatch problem without communications in islanded AC microgrids consisting of cascaded inverters, this paper proposes an optimal decentralized economical-sharing scheme. In proposed scheme, optimal sharing function of the current is applied to generate the reference voltages. And the frequency is used to drive all distributed generators (DGs) synchronize operation in microgrids. When the microgrid is in steady state, DGs share a single common frequency and current in terms of the proposed scheme. Thus the potential advantages of simplicity and decentralized manner are retained. The AC microgrid model has been developed through simulations and experiments to verify the effectiveness and performance of the proposed scheme., Comment: 16 pages, 14 figures

DC microgrids are becoming popular as effective means to integrate various renewable energy resources. Constant power loads (CPLs) may yield instability due to the negative impedance characteristic. This paper analyzes the stability of the DC microgrid in presence of CPL. Distributed generations (DGs) are controlled by using a distributed controller which aims at current sharing and voltage recovery. For simplicity, a reduced order model is derived on the fundamental of neglecting the transient state of the DC/DC converter. The purpose of this paper is to analyze stability conditions and give the suggestions to design control parameters. The stability conditions are obtained by using inertia theorem. Moreover, this paper makes a further detailed research based on the existed theorems. Simulation results are provided to verify the effectiveness and validity of the proposed theorem., Comment: 10 pages, 4 figures