Fault diagnosis of power transformer based on support vector machine with genetic algorithm

Diagnosis of potential faults concealed inside power transformers is the key of ensuring stable electrical power supply to consumers. Support vector machine (SVM) is a new machine learning method based on the statistical learning theory, which is a powerful tool for solving the problem with small sampling, nonlinearity and high dimension. The selection of SVM parameters has an important influence on the classification accuracy of SVM. However, it is very difficult to select appropriate SVM parameters. In this study, support vector machine with genetic algorithm (SVMG) is applied to fault diagnosis of a power transformer, in which genetic algorithm (GA) is used to select appropriate free parameters of SVM. The experimental data from several electric power companies in China are used to illustrate the performance of the proposed SVMG model. The experimental results indicate that the SVMG method can achieve higher diagnostic accuracy than IEC three ratios, normal SVM classifier and artificial neural network.     

An integrated method for finding key suppliers in SCM

Association rule is a widely used data mining technique that searches through an entire data set for rules revealing the nature and frequency of relationships or associations between data entities. Supplier selection is a significant work in supply chain management. Often, there will be thousands of potential suppliers and identifying a subset of these suppliers can be a complex process of determining a satisfactory subset based on a number of factors. In this paper, the supplier selection can be viewed as the problem of mining a large database of shipment. The proposed method incorporates the extended association rule algorithm of data mining with that of set theory to find key suppliers. This research has employed a numerical example for the integrated method to develop suitable supplier clusters. The results show that the method is effective and applicable.     

A wavelet-tree-based watermarking method using distance vector of binary cluster

This paper proposes a wavelet-tree-based watermarking method using distance vector of binary cluster for copyright protection. In the proposed method, wavelet trees are classified into two clusters using the distance vector to denote binary watermark bits. The two smallest wavelet coefficients in a wavelet tree are used to reduce distortion of a watermarked image. The distance vector, which is obtained from the two smallest coefficients of a wavelet tree, is quantized to decrease image distortion. The trees are classified into two clusters so that they exhibit a sufficiently large statistical difference based on the distance vector, which difference is then used for subsequent watermark extraction. We compare the statistical difference and the distance vector of a wavelet tree to decide which watermark bit is embedded in the embedding process. The experimental results show that the watermarked image looks visually identical to the original and the watermark can be effectively extracted upon image processing attacks.     

一种基于广义期望首达时间的形状距离学习算法

形状距离学习是形状匹配框架中引入的后处理步骤, 能够有效改善逐对计算得到的形状间距离.利用期望首达时间分析形状间相似度可能导致距离更新不准确, 针对这一问题提出了一种基于广义期望首达时间 (Generalized mean first-passage time, GMFPT) 的形状距离学习方法.将形状样本集合视作状态空间, 广义期望首达时间表示质点由一个状态转移至指定状态集合所需的平均时间步长, 本文将其视作更新后的形状间距离.通过引入广义期望首达时间, 形状距离学习方法能够有效地分析上下文相关的形状相似度, 显式地挖掘样本空间流形中的最短路径, 并消除冗余上下文形状信息的影响.将所提出的方法应用到不同形状数据集中进行仿真实验, 本文方法比其他方法能够得到更准确的形状检索结果.     

Augmented reality in science laboratories: The effects of augmented reality on university students’ laboratory skills and attitudes toward science laboratories

This study investigated the effects of the use of augmented reality (AR) technologies in science laboratories on university students' laboratory skills and attitudes towards laboratories. A quasi-experimental pre-test/post-test control group design was employed. The participants were 76 first-year university students, aged 18–20 years old. They were assigned to either an experimental or a control group. Qualitative and quantitative data collection tools were used. The experimental results obtained following the 5-week application revealed that the AR technology significantly enhanced the development of the university students' laboratory skills. AR technology both improved the students’ laboratory skills and helped them to build positive attitudes towards physics laboratories. The statements of the students and the instructor regarding other effects of AR technology on science laboratories, both negative and positive, are also discussed.     

Cooperation and profit allocation in two-echelon logistics joint distribution network optimization

Collaborative two-echelon logistics joint distribution network can be organized through a negotiation process via logistics service providers or participants existing in the logistics system, which can effectively reduce the crisscross transportation phenomenon and improve the efficiency of the urban freight transportation system. This study establishes a linear optimization model to minimize the total cost of two-echelon logistics joint distribution network. An improved ant colony optimization algorithm integrated with genetic algorithm is presented to serve customer clustering units and resolve the model formulation by assigning logistics facilities. A two-dimensional colony encoding method is adopted to generate the initial ant colonies. Improved ant colony optimization combines the merits of ant colony optimization algorithm and genetic algorithm with both global and local search capabilities. Finally, an improved Shapley value model based on cooperative game theory and a cooperative mechanism strategy are presented to obtain the optimal profit allocation scheme and sequential coalitions respectively in two-echelon logistics joint distribution network. An empirical study in Guiyang City, China, reveals that the improved ant colony optimization algorithm is superior to the other three methods in terms of the total cost. The improved Shapley value model and monotonic path selection strategy are applied to calculate the best sequential coalition selection strategy. The proposed cooperation and profit allocation approaches provide an effective paradigm for logistics companies to share benefit, achieve win–win situations through the horizontal cooperation, and improve the negotiation power for logistics network optimization.     

On scheduling transactions in a grid processing system considering load through Ant Colony Optimization

Load balanced transaction scheduling problem is an important issue in distributed computing environments including grid system. This problem is known to be NP-hard and can be solved by using heuristic as well as any meta-heuristic method. We ponder over the problem of the load balanced transaction scheduling in a grid processing system by using an Ant Colony Optimization for load balancing. The problem that we consider is to achieve good execution characteristics for a given set of transactions that has to be completed within their given deadline. We propose a transaction processing algorithm based on Ant Colony Optimization (ACO) for load balanced transaction scheduling. We modify two meta-heuristic along with ACO and three heuristic scheduling algorithms for the purpose of comparison with our proposed algorithm. The results of the comparison show that the proposed algorithm provides better results for the load balanced transaction scheduling in the grid processing system.     

Evidential grammars: A compositional approach for scene understanding. Application to multimodal street data

Automatic scene understanding from multimodal data is a key task in the design of fully autonomous vehicles. The theory of belief functions has proved effective for fusing information from several sensors at the superpixel level. Here, we propose a novel framework, called evidential grammars, which extends stochastic grammars by replacing probabilities by belief functions. This framework allows us to fuse local information with prior and contextual information, also modeled as belief functions. The use of belief functions in a compositional model is shown to allow for better representation of the uncertainty on the priors and for greater flexibility of the model. The relevance of our approach is demonstrated on multi-modal traffic scene data from the KITTI benchmark suite.