Computer Development & Applications 

Abstract: A short-term load forecasting method that is based on genetic algorithm(GA) and least squares support vector machines(LS-SVM) is proposed.Because there are various factors impacting the accuracy of load forecasting and uncertain parameters to LS-SVM,the paper proposes to select features and parameters using genetic algorithm.Then,the LS-SVM model is build based on the selected features and parameters.Simulation results show that the algorithm can improve prediction model accuracy and generalization ability.

Abstract: Internet and technology information field in the wide application of education,and the variety distance of teaching,make online exam system arises open.Based on the function requirement analysis and function design to the system,the key technologies is used,such as ASP.NET Microsoft SQL Server 2005 and Windows Server 2003,the user login,online examination,reading scores,question bank management and administrator module was designed,at the same time to achieve the user login,modify password,on-line test and other major modules.

Abstract: Control system of the irradiation electron linac is designed and implemented based on Distributed Control System(DCS). The hardware consists of Siemens Programmable Logical Controllers(PLCs)and industrial computers. The bus between PLCs is MPI and industrial network between computer and the PLCs. The software is developed via Experimental Physics and Industrial Control System. The EPICS IOC connected the PLCs via TCP/IP, using the driver support. The OPI was designed in friendly interface and the archiving system is realized base on My SQL database. The control system operates stably through the test.

Abstract: The performance of image feature points extraction determinces the efficiency of image matchingThis paper proposes a SIFT algorithm based on multiscale spaceThe algorithm has better adaptility for image rotation,scaling zoom and brightness varirentThis paper analyses the basic model of SIFT and the access of key points on location,scale and direction parameter based on scale spaceExperiments show that the algorithm has high accuracy for the access of the feature points information on scale,orientation and size in different scale spaceIt also has strong stability and robustness