基于GRU-RNN组合模型的云计算RLF仿真

随着云计算的兴起和信息技术的快速发展,云计算数据中心存在能源消耗过高的问题,服务器的能源消耗主要是由CPU和内存等资源消耗过高引起的,这些资源利用率的不平衡是导致资源浪费和能源消耗的关键因素之一。对数据中心资源使用量进行预测,有助于进行资源管理,提高服务器资源利用率,从而缓解资源浪费和能耗过高的问题。本文建立一种基于门控循环单元与循环神经网络的组合预测模型GRU-RNN对负载进行预测。实验结果表明,提出的GRU-RNN模型相比于以往简单预测模型RNN、LSTM、GRU和现有的复合预测模型ARIMA-LSTM、GRU-LSTM等,有更高的预测精度。

基于国际金融中心建设目标的上海金融人力资源评估与总量需求预测

金融行业人力资源的存量和发展状况决定了国际金融中心城市的综合竞争力和可持续发展能力。本文用历史数据趋势分析和与香港、纽约等主要标杆国际金融中心城市横向比较的方法,构建了上海金融人力资源需求总量的预测模型。在评估上海金融人力资源现状和与香港、纽约的金融人力资源总量和人均贡献进行国际比较的基础上,预测了上海至2020年建成国际金融中心对金融人力资源总量的需求,计算和分析了金融人才需求的总量和结构缺口。对未来十几年上海金融人力资源建设提供了参考的目标。

An Improved Markov Chain Model Based on Autocorrelation and Entropy Techniques and Its Application to State Prediction of Water Resources

According to the relationships among state transition probability matrixes with different step lengths, an improved Markov chain model based on autocorrelation and entropy techniques was introduced. In the improved Markov chain model, the state transition probability matrixes can be adjusted. The steps of the historical state of the event, which was significantly related to the future state of the event, were determined by the autocorrelation technique, and the impact weights of the event historical state on the event future state were determined by the entropy technique. The presented model was applied to predicting annual precipitation and annual runoff states, showing that the improved model is of higher precision than those existing Markov chain models, and the determination of the state transition probability matrixes and the weights is more reasonable. The physical concepts of the improved model are distinct, and its computation process is simple and direct, thus, the presented model is sufficiently general to be applicable to the prediction problems in hydrology and water resources.

Study on resource quantity of surface water based on phase space reconstruction and neural network

Proposed a new method to disclose the complicated non-linearity structure of the water-resource system, introducing chaos theory into the hydrology and water resources field, and combined with the chaos theory and artificial neural networks. Training data construction and networks structure were determined by the phase space reconstruction, and establishing nonlinear relationship of phase points with neural networks, the forecasting model of the resource quantity of the surface water was brought forward. The keystone of the way and the detailed arithmetic of the network training were given. The example shows that the model has highly forecasting precision.

A Global-Scale Image Lossless Compression Method Based on QTM Pixels

In this paper, a new predictive model, adapted to QTM (Quaternary Triangular Mesh) pixel compression, is introduced. Our approach starts with the principles of proposed predictive models based on available QTM neighbor pixels. An algorithm of ascertaining available QTM neighbors is also proposed. Then, the method for reducing space complexities in the procedure of predicting QTM pixel values is presented. Next, the structure for storing compressed QTM pixel is proposed. In the end, the experiment on comparing compression ratio of this method with other methods is carried out by using three wave bands data of 1 km resolution of NOAA images in China. The results indicate that: 1) the compression method performs better than any other, such as Run Length Coding, Arithmetic Coding, Huffman Cod- ing, etc; 2) the average size of compressed three wave band data based on the neighbor QTM pixel predictive model is 31.58% of the origin space requirements and 67.5% of Arithmetic Coding without predictive model.

Application of a Neural Network Technique for Prediction of the Water Quality Index in the Dong Nai River, Vietnam

Recent trends in environmental management of water resource have enlarged the demand for predicting techniques that can provide reliable, efficient and accurate water quality. In this case study, the authors applied the Artificial Neural Networks （ANN） to estimate the water quality index on the Dong Nai River flowing through Dong Nai and Binh Duong provinces. The information and data including 10 water quality parameters of the Dong Nai River at 23 monitoring stations were collected during the recorded time period from 2010 to 2014 to build water quality prediction models. The results of the study demonstrated that the Water Quality Index （WQI） forecasted with GRNN was very significant and had high correlation coefficient （R2 = 0.974 and p = 0.0） compared to the real values of the WQI. Moreover, the ANN models provided better predicted values than the multiple regression models did.

Pointwise Convergence of a Nonparametric Estimator of Regression in a Measurable Space Used in Contingent Valuation Method

The Contingent Valuation Method is used to evaluate individual preferences for a change concerning a public non-market resource or property. The objective is to build a nonparametric forecasting model of an individual＇s Willingness To Pay according to geographical location. Within this framework, an estimator （of type Nadaraya-Watson） is proposed for the regression of the variable related to geolocation. The specific characteristics of the location variable lead us to a more general regression model than the traditional models. Results are established for convergence of our estimator.

Traffic Characteristics Based Dynamic Radio Resource Management in Heterogeneous Wireless Networks

The traffic with tidal phenomenon in Heterogeneous Wireless Networks(HWNs)has radically increased the complexity of radio resource management and its performance analysis.In this paper,a Simplified Dynamic Hierarchy Resource Management(SDHRM)algorithm exploiting the resources dynamically and intelligently is proposed with the consideration of tidal traffic.In network-level resource allocation,the proposed algorithm first adopts wavelet neural network to forecast the traffic of each sub-area and then allocates the resources to those sub-areas to maximise the network utility.In connection-level network selection,based on the above resource allocation and the pre-defined QoS requirement,three typical network selection policies are provided to assign traffic flow to the most appropriate network.Furthermore,based on multidimensional Markov model,we analyse the performance of SDHRM in HWNs with heavy tailed traffic.Numerical results show that our theoretical values coincide with the simulation results and the SDHRM can improve the resource utilization.

Research on Prediction of Enterprise Human Resource Demand and Supply Based on Gray Model

As the knowledge-intensive enterprise, the core competence of software enterprise is to master human resources of knowledge. But the unique human resource characteristic of software enterprise res^cts the management. While it develops rapidly, there is great waste for human resource cost. By using the prediction model of human resources for software enterprise, the paper proposes the human resource strategic planning system model of software enterprise. The paper applies grey prediction to make empirical research on requirements of human resource for software enterprises, and uses Markov chain to analyze supply of human resources. According to the prediction results, the paper adjusts the imbalanced supply and demand, and proposes the human resources planning.