In order to analyze the influence rule of experimental parameters on the energy-absorption characteristics and effectively forecast energy-absorption characteristic of thin-walled structure, the forecast model of GA-B...In order to analyze the influence rule of experimental parameters on the energy-absorption characteristics and effectively forecast energy-absorption characteristic of thin-walled structure, the forecast model of GA-BP hybrid algorithm was presented by uniting respective applicability of back-propagation artificial neural network （BP-ANN） and genetic algorithm （GA）. The detailed process was as follows. Firstly, the GA trained the best weights and thresholds as the initial values of BP-ANN to initialize the neural network. Then, the BP-ANN after initialization was trained until the errors converged to the required precision. Finally, the network model, which met the requirements after being examined by the test samples, was applied to energy-absorption forecast of thin-walled cylindrical structure impacting. After example analysis, the GA-BP network model was trained until getting the desired network error only by 46 steps, while the single BP-ANN model achieved the same network error by 992 steps, which obviously shows that the GA-BP hybrid algorithm has faster convergence rate. The average relative forecast error （ARE） of the SEA predictive results obtained by GA-BP hybrid algorithm is 1.543%, while the ARE of the SEA predictive results obtained by BP-ANN is 2.950%, which clearly indicates that the forecast precision of the GA-BP hybrid algorithm is higher than that of the BP-ANN.展开更多
For accelerating the supervised learning by the SpikeProp algorithm with the temporal coding paradigm in spiking neural networks (SNNs), three learning rate adaptation methods (heuristic rule, delta-delta rule, and de...For accelerating the supervised learning by the SpikeProp algorithm with the temporal coding paradigm in spiking neural networks (SNNs), three learning rate adaptation methods (heuristic rule, delta-delta rule, and delta-bar-delta rule), which are used to speed up training in artificial neural networks, are used to develop the training algorithms for feedforward SNN. The performance of these algorithms is investigated by four experiments: classical XOR (exclusive or) problem, Iris dataset, fault diagnosis in the Tennessee Eastman process, and Poisson trains of discrete spikes. The results demonstrate that all the three learning rate adaptation methods are able to speed up convergence of SNN compared with the original SpikeProp algorithm. Furthermore, if the adaptive learning rate is used in combination with the momentum term, the two modifications will balance each other in a beneficial way to accomplish rapid and steady convergence. In the three learning rate adaptation methods, delta-bar-delta rule performs the best. The delta-bar-delta method with momentum has the fastest convergence rate, the greatest stability of training process, and the maximum accuracy of network learning. The proposed algorithms in this paper are simple and efficient, and consequently valuable for practical applications of SNN.展开更多
Used genetic algorithm (GA) to optimize the network of ventilation in order toavoid artificial convergence and speed up the convergence rate to introduce the Powellalgorithm. The Powell algorithm had been integrated i...Used genetic algorithm (GA) to optimize the network of ventilation in order toavoid artificial convergence and speed up the convergence rate to introduce the Powellalgorithm. The Powell algorithm had been integrated into GA. Powell had the effectivecapacity of solving the local optimal solution. Powell and the cross as a method ofchoice, a variation of the parallel operator, can be a better solution to the prematureconvergence of the GA problem. The two methods will be improved to make it an effective combination of hybrid GA called hybrid genetic algorithm (HGA) for the introductionof mine ventilation network optimization and to be used to solve the problem of regulating mine optimization.展开更多
基金Project(50175110) supported by the National Natural Science Foundation of ChinaProject(2009bsxt019) supported by the Graduate Degree Thesis Innovation Foundation of Central South University, China
文摘In order to analyze the influence rule of experimental parameters on the energy-absorption characteristics and effectively forecast energy-absorption characteristic of thin-walled structure, the forecast model of GA-BP hybrid algorithm was presented by uniting respective applicability of back-propagation artificial neural network （BP-ANN） and genetic algorithm （GA）. The detailed process was as follows. Firstly, the GA trained the best weights and thresholds as the initial values of BP-ANN to initialize the neural network. Then, the BP-ANN after initialization was trained until the errors converged to the required precision. Finally, the network model, which met the requirements after being examined by the test samples, was applied to energy-absorption forecast of thin-walled cylindrical structure impacting. After example analysis, the GA-BP network model was trained until getting the desired network error only by 46 steps, while the single BP-ANN model achieved the same network error by 992 steps, which obviously shows that the GA-BP hybrid algorithm has faster convergence rate. The average relative forecast error （ARE） of the SEA predictive results obtained by GA-BP hybrid algorithm is 1.543%, while the ARE of the SEA predictive results obtained by BP-ANN is 2.950%, which clearly indicates that the forecast precision of the GA-BP hybrid algorithm is higher than that of the BP-ANN.
基金Supported by the National Natural Science Foundation of China (60904018, 61203040)the Natural Science Foundation of Fujian Province of China (2009J05147, 2011J01352)+1 种基金the Foundation for Distinguished Young Scholars of Higher Education of Fujian Province of China (JA10004)the Science Research Foundation of Huaqiao University (09BS617)
文摘For accelerating the supervised learning by the SpikeProp algorithm with the temporal coding paradigm in spiking neural networks (SNNs), three learning rate adaptation methods (heuristic rule, delta-delta rule, and delta-bar-delta rule), which are used to speed up training in artificial neural networks, are used to develop the training algorithms for feedforward SNN. The performance of these algorithms is investigated by four experiments: classical XOR (exclusive or) problem, Iris dataset, fault diagnosis in the Tennessee Eastman process, and Poisson trains of discrete spikes. The results demonstrate that all the three learning rate adaptation methods are able to speed up convergence of SNN compared with the original SpikeProp algorithm. Furthermore, if the adaptive learning rate is used in combination with the momentum term, the two modifications will balance each other in a beneficial way to accomplish rapid and steady convergence. In the three learning rate adaptation methods, delta-bar-delta rule performs the best. The delta-bar-delta method with momentum has the fastest convergence rate, the greatest stability of training process, and the maximum accuracy of network learning. The proposed algorithms in this paper are simple and efficient, and consequently valuable for practical applications of SNN.
基金Supported by the National Natural Science Foundation of China(60772159)
文摘Used genetic algorithm (GA) to optimize the network of ventilation in order toavoid artificial convergence and speed up the convergence rate to introduce the Powellalgorithm. The Powell algorithm had been integrated into GA. Powell had the effectivecapacity of solving the local optimal solution. Powell and the cross as a method ofchoice, a variation of the parallel operator, can be a better solution to the prematureconvergence of the GA problem. The two methods will be improved to make it an effective combination of hybrid GA called hybrid genetic algorithm (HGA) for the introductionof mine ventilation network optimization and to be used to solve the problem of regulating mine optimization.
