Focused on the issue that the traditional interest area based visualization method can not pay attention to the details in the process of analyzing pilot eye movement data, a visual analysis method of eye movement data based on user-defined interest area was proposed. Firstly, according to the specific analysis task, the self-divison and self-definition of the background image of the task were introduced. Then, multiple auxiliary views and interactive approaches were combined, and an eye movement data visual analysis system for pilot training was designed and implemented to help analysts analyze the difference of eye movement between different pilots. Finally, through case analysis, the effectiveness of the visual analysis method and the practicability of the analysis system were proved. The experimental results show that compared with the traditional method, in the proposed method, the analysts' initiative in the analysis process is increased. The analysts are allowed to explore the local details of the task background in both global and local aspects, making the analysts' analyze the data in multi-angle; the analysts are allowed find the flight students' cognitive difficulties in the training process as a whole, so as to develop more targeted and more effective training courses.

Since New Approach to Multi-Objective A ^* combined with dimensionality reduction technique (NAMOA _dr ^*) algorithm has the phenomenon of plateau exploration, a Random Walk assisted NAMOA _dr ^* (RWNAMOA _dr ^*) algorithm which invoked a random walk procedure was proposed to find an exit (labels with heuristic value not dominated by the last extended label's) when the NAMOA _dr ^*was stuck on a plateau. To determine when NAMOA _dr ^* algorithm was stuck on a plateau exploration, a method of detecting plateau exploration was proposed. When the heuristic value of the extended label was dominated by the last extended label's for continuous m times, NAMOA _dr ^* algorithm was considered to fall into the plateau exploration. In the experiments, a randomly generated grid was used, which was a standard test platform for the evaluation of multi-objective search algorithms. The experimental results reveal that compared with NAMOA _dr ^* algorithm, RWNAMOA _dr ^* algorithm's running time is reduced by 50.69% averagely and its space consuming is reduced by about 10% averagely, which can provide theoretical support for accelerating multi-objective path searching in real life.

The α-shape algorithm is not suitable for surface reconstruction of scattered and non-uniformly sampled points. In order to solve the problem, an improved surface reconstruction algorithm with adaptive α-shape based on Local Feature Size (LFS) of point cloud data was proposed. Firstly, Medial Axis (MA) of the surface was approximated by the negative poles computed by k-nearest neighbors of sampled points. Secondly, the LFS of sampled points was calculated by the approximated MA, and the original point clouds were unequally simplified based on LFS. Finally, the surface was adaptively reconstructed based on the radius of circumscribed ball of triangles and the corresponding α value. In the comparison experiments with α-shape algorithm, the proposed algorithm could effectively and reasonably reduce the number of point clouds, and the simplification rate of point clouds achieved about 70%. Simultaneously, the reconstruction result were obtained with less redundant triangles and few holes. The experimental results show that the proposed algorithm can adaptively reconstruct the surface of non-uniformly sampled point clouds.

For the problem of optimizing resource allocation to achieve profit maximization of cloud computing center, an analysis model based on Service Layer Agreement (SLA)-aware was proposed for optimizing server number and speed of cloud center. Meanwhile some important factors were taken into account, such as energy cost, server rental cost, customer waiting time, and SLA violation penalty. The impacts of cloud center profit by changing server number and speed were analyzed by numerical simulation. The numerical simulation results indicate that cloud center will obtain maximum profit by optimizing server number and speed at a certain request rate; with request rate increasing, profit will increase linearly by optimizing server number and speed. The analysis results can provide a reference method for cloud service provider to improve net business gain.

null

For the problem of minimizing the energy consumption under performance constraint of cloud center, an optimal power consumption allocation method among multiple heterogeneous servers was proposed. First, an optimal energy consumption mathematical model of cloud center was built. Second, a Minimizing Power Consumption (MPC) algorithm for calculating the minimum energy was developed by using Lagrange multiplier method to obtain the optimal solution of the model. Finally, the MPC algorithm was verified by plenty of numerical experiments and compared with the Equal-Power (EP) baseline method. The experimental results indicate that MPC algorithm can save approximately 30% energy than the EP baseline method under the same load and the same response time conditions, and the proportion of energy saving will increase with load increasing. The MPC algorithm can effectively avoid energy configuration overload and it will provide ideas and reference data for optimal resource allocation of cloud center.

Since it is necessary to evaluate and analyze the service performance of cloud computing center to guarantee Quality of Service (QoS) and avoid violation of Service Layer Agreement (SLA), a approximated analysis model based on M/M/n/n+r queue theory was proposed for cloud computing center. By solving this model the probability distribution function of response time and other QoS indicators were acquired, meanwhile the relationship among the number of servers, size of queue buffers, response time, blocking probability and instance service probability were revealed and verified by simulation.The experimental results indicate that improving server service rate is better than increasing the number of servers for improving service performance.

Concerning the problem that previous studies on the scalability do not fully consider parallel execution time, and the relationships between latency scalability and parallel execution time have not been yet studied thoroughly, this paper studied the relationships between latency scalability and parallel execution time deeply and fully. Thereby some important conclusions were drawn, and they were about the relationships between latency scalability and parallel execution time after different algorithm-machines were extended from the same initial state. Then the proof of the above conclusions was given in this paper. The derived conclusions enriched the research content about the relationships between latency scalability and parallel execution time and provided a theoretical basis for obtaining ideal latency scalability of parallel computing. Finally the important conclusions and analytical expressions were verified through experimental results obtained for different algorithm-machines.

Recently, Extreme Learning Machine (ELM) has been a promising tool in solving a wide range of classification and regression problems. However, the generalization performance of ELM will be decreased when there exits redundant hidden neurons. To solve this problem, this paper introduced a new regularizer that was the Frobenius norm of mapping matrix from hidden space to a new latent feature space. Furthermore, an alternating optimization strategy was adopted to learn the above regularization problem and the latent feature space. The proposed algorithm was tested empirically on the classical UCI data set as well as a load identification engineering data set. The experimental results show that the proposed algorithm obviously outperforms the classical ELM in terms of predictive precision and numerical stability, and needs much less computational cost than the present ELM model selection algorithm.

To solve the slow rendering speed issue of serial photon mapping algorithm, the feasibility of parallelizing the algorithm was analyzed. The parallelism and computing capability of the Compute Unified Device Architecture (CUDA) on Graphic Processing Unit (GPU) were fully utilized to realize a parallel photon mapping algorithm. As for the shortage of generating the same number of GPU threads as the photon number in the photon emission and tracing step, and the waste of resources of the average allocation method, a new cooperation way that all the threads be processed with dynamic balance was then proposed. The new method nearly doubled the rendering speed. The experimental results prove the effectiveness of the proposed method.

Considering the data security and integrity on the processing of HD video data stream transmission, acrypto DSP with special implementation for DES, SHA1, MD5, RSA are introduced. In order to improve the performance and decrease the cost, the DSP has 11 pipeline stages, and two parallel execution clusters (each cluster contains 3 function units). In order to improve throughput, special instructions are customized for complex operations. The methods of realizing the symmetrical encryption, public-key encryption and authentication algorithm based on such DSP are presented. In order to improve the throughput, The simulation experiment results show that the performance can well satisfy the requirement of real time HD video data stream applications.

By analyzing the principle and the existing problems in the ray method and the extreme coordinate value method, the existing algorithms which determined contour direction were simplified. Then an improved algorithm on the contour line position relation was proposed combining the advantages of the ray method and the extreme coordinate method. The algorithm mainly included four parts: distinction among the internal and external contours, adjustment of the profile direction, inclusive identification of contours and the construction of a contour tree. The experimental results show that the improved algorithm can correct the problems in the existing algorithms and achieve good efficiency.

Concerning the problem of large computation, slow response and repeated computation in travel sky-based Calendar Shopping (CS) system, an efficient method was proposed to cache calculation results of unit in this paper. The system architecture was redesigned and the performance of calendar shopping system was optimized. The experimental results show that the presented method can reduce the system response time and improve the performance of the system significantly, also it provides the method and theoretical support for calendar shopping system in civil aviation field.

Ad Hoc network often supplies reliable authentication service for nodes by distributed certificate authority authentication model due to its dynamic topology and other characteristics; however, the problem of secure auditing is not solved by existing schemes. Therefore, a distributed secret share distribution scheme was put forward based on threshold warrant certificates, which carried out strict auditing towards the nodes that applied for secret shares, and could effectively withstand several malicious nodes working together to recover the secret key of system, and make sure that only the credible and high-quality-service nodes were able to gain the secret shares. In the end, the security and success probability of the scheme were analyzed in theory, and its effectiveness was confirmed by simulations.

After discussing the debugging and emulation of traditional ARM emulators,  a perfect design of JTAG emulator for ARM over Ethernet was presented, based on the central CPU SAMSUNG S3C44B0X. This ARM emulator over Ethernet, can provide high downloading speed,  and shared the emulator resouces.