Applied Mechanics and Materials Vols. 373-375

Abstract: Base on improved particle swarm algorithm, this paper proposes a linear decreasing inertia weight particle swarm algorithm and error back propagation algorithm based on hybrid algorithm combining. The linear decreasing inertia weight particle swarm algorithm and momentum-adaptive learning rate BP algorithm interchangeably adjust the network weights, so that the two algorithms are complementary. It gives full play to the PSO's global optimization ability and the BP algorithm local search advantage, to overcome the slow convergence speed and easily falling into local weight problems. Simulation results show that this diagnostic method can be used for tolerance analog circuit fault diagnosis, with a high convergence rate and diagnostic accuracy.

Abstract: Kernel parameter selection of support vector machine (SVM) is difficult in practical application. A parameter selection algorithm of SVM was proposed based on data maximum variance - entropy criterion by analyzing the principle of SVM classifier. The algorithm uses data maximum variance - entropy criterion to measure the linear separability of dataset in the feature space, and combines with particle swarm optimization (PSO) algorithm for parameter optimization. The experiment results on datasets from UCI show that the algorithm is excellence in accuracy and improves the training performance of SVM. To further verify the effectiveness of the algorithm, applying the method in fault diagnosis of biquadratic filter circuit, results prove it improves the diagnostic accuracy.

Abstract: The integration of variable precision rough set and neural network is introduced into the bearing fault diagnosis. VPRS-INN fault diagnosis method is proposed: First, utilize the information entropy method for discretization of continuous attributes, and then use attribute dependence degree of the variable precision rough set theory for heuristic reduction. based on the reduction, obtain the optimal decision support system. Finally according to the optimal design system, we design a integrated neural network for fault diagnosis. instances have proved the feasibility and high fault diagnosis rate of the method.

Abstract: When a submarine uses an anti-torpedo tactic, it is a matter of life or death. There are two types of countermeasures:decoys and jammers. A successful anti-torpedo tactic should consist of the deployment of mixed decoys and jammers and the coordination with the submarines maneuver. This paper would like to discuss the anti-torpedo tactics and study the interaction among the submarine and torpedo. After applying the evolutionary algorithm, it is interesting to discover the survivability of submarine in the torpedoes engagement scenario.

Abstract: Suspension spring is used in the suspension system of light vehicle and medium buses widely, and its design quality related to stability and security of the vehicle. This paper take the suspension coil spring of a light vehicle as the research object, its multi-objective optimization model is established. The volume of spring and one frequency free vibration frequency are taken as optimization objective, the strength, stiffness, stability, fatigue strength and the winding ratio of the spring are taken as constraints, and use NSGA-II algorithm, obtained Pareto optimal solution set of the optimization problem. The coil spring model and optimization method used in this paper is also suitable for optimization design of other spring.

Abstract: Particle swarm optimization algorithms have lots of advantages such as fast convergence speed, good quality of solution and robustness in multidimensional space function optimization and dynamic target optimization. It is suitable for structural optimization design. In this paper, manual transmission gear train of a tractor is taken as research object, the minimum quality and minimum center distance of the gear train is taken as optimization goal, the gear ratio, modulus, helix angle, tooth width and equilibrium conditions of the axial force are taken as the constraints, a multi-objective optimization model of the gear train is established. The optimal structure design programs and Pareto optimal solution are obtained by using particle swarm optimization algorithm.

Abstract: The paper gave a new frequent item sets mining algorithm based on index table at multiple times for the Apriori algorithm scans the database which causes the I/O load is too large, and the costly problem with the Apriori algorithm which want to have a big candidate sets. The algorithm first generated a one-dimensional index table by scan the database once, and then generates a two-dimensional index table based on the one-dimensional index table. After the two-dimension index table had been generated, we can use the method similar with Floyd algorithm, which inserts the single index entry individually into the two-dimensional index table. If the count of new index value is greater than or equal to Minsuppor after the single index item had been inserted, the new index entrys Item will be a frequently item sets. After all single index entry had been inserted into the two-dimensional index table, all the index entry in the table will be the maximum frequently item sets. After analysis we can see that this algorithm has low cost and with the high accuracy than Apriori algorithm and can provide some reference for related rules.

Abstract: This paper expounds diversified methods for hardware resources accessing and prot data reading and writing in MATLAB/Simulink, as well the diversified methods are analyzed. It provides a more direct and more effective platform for achieving real-time and on-line measure & control in MATLAB and studying on advanced control laws.

Abstract: A improved binary tree SVM multi-class classification algorithm is proposed. Firstly, constructing the minimum hyper ellipsoid for each class sample in the feather space, and then generating optimal binary tree according to the hyper ellipsoid volume, training sub-classifier for every non-leaf node in the binary tree at the same time. For the sample to be classified, the sub-classifiers are used from the root node until one leaf node, and the corresponding class of the leaf node is the class of the sample. The experiments are done on the Statlog database, and the experimental results show that the algorithm improves classification precision and classification speed, especially in the situation that the number of class are more and their distribution area are equal approximately, the algorithm can greatly improve the classification precision and classification speed.

Abstract: Estimation of distribution algorithms (EDAs) is a method for solving NP-hard problem. But it is hard to find global optimization quickly for some problems, especially for traveling salesman problem (TSP) that is a classical NP-hard combinatorial optimization problem. To solve TSP effectively, a novel estimation of distribution algorithm (NEDA ) is provided, which can solve the conflict between population diversity and algorithm convergence. The experimental results show that the performance of NEDA is effective.