Advanced Materials Research Vols. 971-973

Abstract: In recent years, a lot of business and industry rely on a large amount of data. In some scenarios, it is important to be able to predict new data based on existing data, which is known as the prediction problem. We want the prediction to be simple and accurate. Artificial neural network is a very popular model for learning linear and nonlinear functions. In this paper, we propose to use artificial neural network to do intelligent data prediction. According to our experimental results, our proposed method is very accurate.

Abstract: The modern B2B electronic business system should be able to make the enterprise meet better the need of the market, which must have some characteristics of cross-organization, cross-enterprises and frequent interaction. The paper proposes a dynamic electronic commerce model based on XML and Web Services, which consists of presentation layer, Web Services layer, application service layer, data integration layer and data layer. The dynamic e-business model can not only implement the loose system integration, but also be able to reuse the existing resources, and reduce to the cost of system integration and implementation time.

Abstract: The edge detection which comes from the classical Sobel operator in the image is based on the horizontal gradient and vertical gradient direction. On the basis of the template from two directions the method which is to detect the edges of eight directions is discussed in this paper. The image edge detection based on multiple directions can improve the accuracy of edge detection. The numerical experimental results show that the proposed method can get the smooth, continuous, multiple directions edges, and can reduce the loss of the information. The method proposed in this paper is to detect the edge of the image, which will result in higher accuracy. Especially for more complex texture image, the effect of edge detection is more obvious, and the pixel width is closer to a single pixel width. It is an effective method for edge detection.

Abstract: For more effectively solving SDCP,in the paper,using BP neural networks to approximate chance function,training samples are produced by random simulation,and a hybrid intelligent algorithm for SDCP combined stochastic particle swarm optimization and BP neural network is proposed.The experimental results show that the algorithm is more preferable.

Abstract: The distribution center is the bridge connected supply points and demand points, lies in pivotal status in modern logistics system. Firstly, the mathematical model of a distribution center location is established, based on the study of using neural network to solve distribution center location of the previous scholars, a new method is presented as well as the improved immune algorithm. A new affinity formula is designed for immunoselection criteria. Simultaneously, based on the mathematical model and cases, the algorithm is drilled concrete. A case shows that the improved immune algorithm can better solve the problem of the distribution center location.

Abstract: The main purpose of designing an evaluation method for routing strategy is to accurately evaluate the performance of a series of routing strategies at various network loads, including under-load, full- load and even overload. Most researchers evaluate the effectiveness of a new routing strategy with the method of Order Parameter. However in this evaluating process, we find that its result distributes in a certain range of values instead of a single value, so Order Parameter introduces a considerable error, it is difficult to evaluate the performance of a routing strategy accurately in different network load with it. To solve this problem, a method--Improved Order Parameter (OOP) is proposed. This method can not only get a fixed critical value for a special network load, but also depict the transition of network load over time. According to simulation experiment, the OOP can be more objective to evaluate the performance of a routing strategy at various network load.

Abstract: Due to the fluctuation of RSSI (Received Signal Strength Indicator) value, the positioning accuracy of indoor localization algorithm that based on RSSI value is not high and unstable, RSSI filtering and reference distance used in this paper. Fluctuation amplitude of RSSI is reduced by mean filtering, and a known distance between anchor nodes introduced when calculating the unknown distance between blind node and anchor node. According to the result of experiment, the positioning accuracy of new localization algorithm can reach 1.45 meter, and it can meet the demand of indoor positioning.

Abstract: Type of computer animation using 3D technology can be divided under three seeds that mimic 2D animation, imitating traditional 3D animation, imitating real movie. Mimic 2D animation and imitating real movie are infinitely close with 2D animation and real film movie. Imitating traditional 3D animation is completely different from the traditional three-dimensional animation, which won the visual realism. "Visual realism" to get that 3D animation and real film movie stood on the same competition platform.

Abstract: In this paper, we propose an IPTV controllable multicast solution, which combines the technology of SCB, IGMP and the extended OAM. SCB and IGMP are responsible for the distribution of multicast service. A method that based on the extended OAM multicast control packets is used to realize the functions of control and management in multicast service. We also explain the bandwidth design methods of the EPON system. In the end, the implementation scheme of EPON controllable multicast in IPTV application is proposed.

Abstract: the power spectrum estimation researches various characteristics of signals in the frequency domain. The purpose is that signals are recognized and extract .Because these useful signals are submerged in noise .The article introduces estimation in the classic power spectrum and modern power spectrum. It is important that algorithm of AR model parameters are introduced in the parameter estimation of several typical. It discusses the advantages and disadvantages of various algorithms, and with the help of MATLAB platform, the various algorithms of power spectrum are simulated, in order to undertake choosing according to the actual situation.