Applied Mechanics and Materials Vols. 239-240

Abstract: A new particle swarm optimization (PSO) algorithm (a PSO with Variety Factor, VFPSO) based on the PSO was proposed. Compared with the previous algorithm, the proposed algorithm is to update the Variety Factor and to improve the inertia weight of the PSO. The target of the improvement is that the new algorithm could go on enhancing the robustness as before and should reduce the risk of premature convergence. The simulation experiments show that it has great advantages of convergence property over some other modified PSO algorithms, and also avoids algorithm being trapped in local minimum effectively. So it can avoid the phenomenon of premature convergence.

Abstract: PrefixSpan algorithm will construct huge amount of projected databases in the process of mining sequence patterns, especially mining dense dataset and long sequence pattern, which will cause decline of the performance of the algorithm. The resource problem can be solved by Projection position-based Sequential Pattern Mining Algorithm so as to reduce time and storage space. In order to avoid producing huge amount of projected databases and reduce unnecessary storage space and scanning time, compared with the others improved algorithm, the PSPM utilizes projected position to locate projected sequence position for mining local frequency items and deletes the non-frequent items. Experiment results demonstrate that PSPM outperforms the PrefixSpan(with pseudo-pro) algorithm in the aspect of Runtime performance.

Abstract: This thesis aims to discuss the problems now existing in mining probabilistic of MapReduce Apriori, and to put forward an algorithm about mining probabilistic frequent itemsets based on cloud computing. This algorithm proposes to reduce the quantity of candidate item sets by the strategy of designing to decrease candidate item sets, and divide data sets and candidate item sets into related nodes in order to minimize candidate item sets. By compressing the transaction set to accomplish connecting optimization, it can avoid producing massive alternatives item sets in the process of self-link to improve the operation efficiency of algorithm.

Abstract: The relevance vector machine (RVM) was a Bayesian framework for learning sparse regression models and classifiers, it used single kernel function to map training data from low dimension sample space to high dimension feature space. The prediction accuracy and generalization of traditional single-kernel RVM (sRVM) were not ideal both in classification and regression, so we constructed homogeneous and heterogeneous multiple kernels function (MKF) by kernel function combination in which we testified the validity of basic kernel function (BKF) and its parameters we employed by kernel alignment (KA), then we acquired optimized multiple-kernel RVM (mRVM). Experiment results on LIBSVM datasets not only indicate that both homogeneous and heterogeneous multiple-kernel RVM we constructed possess lower error rate in classification and smaller root mean square (RMS) in regression than single-kernel RVM, but also prove the effectiveness of kernel alignment.

Abstract: Through the arduous study of the famous Cohen-Sutherland algorithm, this paper proposes a novel improved algorithm by analyzing its main flaw in efficiency. The line segment without the clipping window will be rejected in our algorithm. By means of adding auxiliary straight-line, this algorithm efficiently calculates intersection points between line and the boundary of clipping window. Thus, this improved algorithm is simple, improving its clipping efficiency and reaching the desired results.

Abstract: Text clustering is one of the most popular topic detection techniques. However, the existing text clustering approaches require that each document has to be partitioned to one and only one cluster. This is not reasonable in some cases for there exist some documents which should not used to constitute topics. This paper firstly models a text document set as a network and designs a method for decomposing such a network, and then proposes a truly original text clustering algorithm for topic detection, called a network decomposition-based text clustering algorithm for topic detection (NDTCATD). The proposed algorithm ensures that meaningless documents can not be used to constitute topics. Experimental results show that NDTCATD is much better than bisecting k-means algorithm in terms of overall similarity and average cluster similarity. Therefore the proposed algorithm is reasonable and effective and is especially suitable for topic detection.

Abstract: To solve anycast routing problem with multiple QoS constraints, a improved hybrid algorithm which combines genetic algorithm and ant colony algorithm is proposed. In the initial period of hybrid algorithm, genetic algorithm was used to distribute pheromones in links and code and optimize control parameters of ant colony algorithm. Through judgment function, this algorithm can judge the time to combine the genetic algorithm with ant colony algorithm, and initialize the pheromones and start the ant colony algorithm at the last period of hybrid algorithm. To avoid hybrid algorithm falling into local optimal solution, a mutation operator was introduced in algorithm hybrid to update local pheromones of new path produced by mutation operation and reduced pheromones concentration on optimal path in time. The NS2 simulation results show that this algorithm can commendably solve the anycast routing problem with multiple QoS constraints, and its performance is better than other two algorithms.

Abstract: To solve the problem of grid task scheduling, an improved algorithm based on particle swarm optimization (PSO) is presented. This paper introduces mutation into PSO. Mutation makes the algorithm jump out local optimization. To some extent, it overcomes the inherent flaw of PSO that falling into local optimization. This paper also introduces trust mechanism into the algorithm to improve the service performance of grid system. The result of simulation experiment shows that the improved algorithm not only makes the complete time minimum, but also have more tasks executed successfully.

Abstract: The image fusion algorithm discussed in this paper which utilizes wavelet decomposition and fuzzy reasoning combines images from diverse imaging sensors into a single composite image. It first decomposed source images through wavelet transform, computed the extent of each source image’s contribution through fuzzy reasoning using the area feature of source images, and then fused the coefficients through weighted averaging with the extents of each source images’ contributions as the weight coefficients. Experimental results indicate the final composite image may have more complete information content or better perceptual quality than any one of the source images.

Abstract: Firstly, according to the situation of multi aircrafts and multi devices, the maintenance support model was established. Secondly, the genetic algorithm flow chart and the coding method were presented, and the fitness function, the selection, the mutations and the crossover were all given out. Finally, the numerical example was given out. The result shows the proposed algorithm yields visual support program for aircrafts and devices, and the applicable and validity of the algorithm are demonstrated.