Forward-backward algorithm, used by watermark decoder for correcting non-binary synchronization errors, requires to traverse a very large scale trellis in order to achieve the proper posterior probability, leading to ...Forward-backward algorithm, used by watermark decoder for correcting non-binary synchronization errors, requires to traverse a very large scale trellis in order to achieve the proper posterior probability, leading to high computational complexity. In order to reduce the number of the states involved in the computation, an adaptive pruning method for the trellis is proposed. In this scheme, we prune the states which have the low forward-backward quantities below a carefully-chosen threshold. Thus, a wandering trellis with much less states is achieved, which contains most of the states with quite high probability. Simulation results reveal that, with the proper scaling factor, significant complexity reduction in the forward-backward algorithm is achieved at the expense of slight performance degradation.展开更多
Many applications require the solution of large nonsymmetric linear systems with multiple right hand sides. Instead of applying an iterative method to each of these systems individually, it is often more efficient to...Many applications require the solution of large nonsymmetric linear systems with multiple right hand sides. Instead of applying an iterative method to each of these systems individually, it is often more efficient to use a block version of the method that generates iterates for all the systems simultaneously. In this paper, we propose a block version of generalized minimum backward (GMBACK) for solving large multiple nonsymmetric linear systems. The new method employs the block Arnoldi process to construct a basis for the Krylov subspace K m(A, R 0) and seeks X m∈X 0+K m(A, R 0) to minimize the norm of the perturbation to the data given in A.展开更多
Performance parameter prediction technology is the core research content of aeroengine health management,and more and more machine learning algorithms have been applied in the field.Regularized extreme learning machin...Performance parameter prediction technology is the core research content of aeroengine health management,and more and more machine learning algorithms have been applied in the field.Regularized extreme learning machine(RELM)is one of them.However,the regularization parameter determination of RELM consumes computational resources,which makes it unsuitable in the field of aeroengine performance parameter prediction with a large amount of data.This paper uses the forward and backward segmentation(FBS)algorithms to improve the RELM performance,and introduces an adaptive step size determination method and an improved solution mechanism to obtain a new machine learning algorithm.While maintaining good generalization,the new algorithm is not sensitive to regularization parameters,which greatly saves computing resources.The experimental results on the public data sets prove the above conclusions.Finally,the new algorithm is applied to the prediction of aero-engine performance parameters,and the excellent prediction performance is achieved.展开更多
基金supported in part by National Natural Science Foundation of China (61101114, 61671324) the Program for New Century Excellent Talents in University (NCET-12-0401)
文摘Forward-backward algorithm, used by watermark decoder for correcting non-binary synchronization errors, requires to traverse a very large scale trellis in order to achieve the proper posterior probability, leading to high computational complexity. In order to reduce the number of the states involved in the computation, an adaptive pruning method for the trellis is proposed. In this scheme, we prune the states which have the low forward-backward quantities below a carefully-chosen threshold. Thus, a wandering trellis with much less states is achieved, which contains most of the states with quite high probability. Simulation results reveal that, with the proper scaling factor, significant complexity reduction in the forward-backward algorithm is achieved at the expense of slight performance degradation.
文摘Many applications require the solution of large nonsymmetric linear systems with multiple right hand sides. Instead of applying an iterative method to each of these systems individually, it is often more efficient to use a block version of the method that generates iterates for all the systems simultaneously. In this paper, we propose a block version of generalized minimum backward (GMBACK) for solving large multiple nonsymmetric linear systems. The new method employs the block Arnoldi process to construct a basis for the Krylov subspace K m(A, R 0) and seeks X m∈X 0+K m(A, R 0) to minimize the norm of the perturbation to the data given in A.
文摘Performance parameter prediction technology is the core research content of aeroengine health management,and more and more machine learning algorithms have been applied in the field.Regularized extreme learning machine(RELM)is one of them.However,the regularization parameter determination of RELM consumes computational resources,which makes it unsuitable in the field of aeroengine performance parameter prediction with a large amount of data.This paper uses the forward and backward segmentation(FBS)algorithms to improve the RELM performance,and introduces an adaptive step size determination method and an improved solution mechanism to obtain a new machine learning algorithm.While maintaining good generalization,the new algorithm is not sensitive to regularization parameters,which greatly saves computing resources.The experimental results on the public data sets prove the above conclusions.Finally,the new algorithm is applied to the prediction of aero-engine performance parameters,and the excellent prediction performance is achieved.
