The basic principle of interval arithmetic and the basic algorithm of the interval Newton methods are introduced.The prototype algorithm can not find any zero in an interval that has zero sometimes,that is,it is insta...The basic principle of interval arithmetic and the basic algorithm of the interval Newton methods are introduced.The prototype algorithm can not find any zero in an interval that has zero sometimes,that is,it is instable.So the prototype relaxation procedure is improved in this paper.Additionally,an immediate test of the existence of a solution following branch_and_bound is proposed,which avoids unwanted computations in those intervals that have no solution.The numerical results demonstrat that the improved interval Newton method is superior to prototype algorithm in terms of solution quality,stability and convergent speed.展开更多
The penalty equation of LCP is transformed into the absolute value equation, and then the existence of solutions for the penalty equation is proved by the regularity of the interval matrix. We propose a generalized Ne...The penalty equation of LCP is transformed into the absolute value equation, and then the existence of solutions for the penalty equation is proved by the regularity of the interval matrix. We propose a generalized Newton method for solving the linear complementarity problem with the regular interval matrix based on the nonlinear penalized equation. Further, we prove that this method is convergent. Numerical experiments are presented to show that the generalized Newton method is effective.展开更多
A Newton iteration-based interval uncertainty analysis method(NI-IUAM) is proposed to analyze the propagating effect of interval uncertainty in multidisciplinary systems. NI-IUAM decomposes one multidisciplinary syste...A Newton iteration-based interval uncertainty analysis method(NI-IUAM) is proposed to analyze the propagating effect of interval uncertainty in multidisciplinary systems. NI-IUAM decomposes one multidisciplinary system into single disciplines and utilizes a Newton iteration equation to obtain the upper and lower bounds of coupled state variables at each iterative step.NI-IUAM only needs to determine the bounds of uncertain parameters and does not require specific distribution formats. In this way, NI-IUAM may greatly reduce the necessity for raw data. In addition, NI-IUAM can accelerate the convergence process as a result of the super-linear convergence of Newton iteration. The applicability of the proposed method is discussed, in particular that solutions obtained in each discipline must be compatible in multidisciplinary systems. The validity and efficiency of NI-IUAM is demonstrated by both numerical and engineering examples.展开更多
This paper presents a hybrid model reliability analysis method based on the damped Newton method with both random and interval variables to solve the hybrid structure reliability problem.The method combines an outer i...This paper presents a hybrid model reliability analysis method based on the damped Newton method with both random and interval variables to solve the hybrid structure reliability problem.The method combines an outer iterative solution and inner layer numerical calculation.In the outer iteration,the method seeks an optimized solution to the interval variable iterative by adding the boundary constraint condition based on the damped Newton optimization theory.In the inner layer solution,the method first reduces the dimension of the random variable through the dimension reduction method,then obtains the first four-order central moment of the function through the application of the Taylor expansion method,and finally calculates the reliability index of the structure according to the fourth-order moment calculation structure of the function.The results of a numerical example and an engineering ten-rod truss structure show that the proposed method can effectively solve the random-interval hybrid reliability problem and has better calculation accuracy than that of the two-layer iterative method.展开更多
文摘The basic principle of interval arithmetic and the basic algorithm of the interval Newton methods are introduced.The prototype algorithm can not find any zero in an interval that has zero sometimes,that is,it is instable.So the prototype relaxation procedure is improved in this paper.Additionally,an immediate test of the existence of a solution following branch_and_bound is proposed,which avoids unwanted computations in those intervals that have no solution.The numerical results demonstrat that the improved interval Newton method is superior to prototype algorithm in terms of solution quality,stability and convergent speed.
文摘The penalty equation of LCP is transformed into the absolute value equation, and then the existence of solutions for the penalty equation is proved by the regularity of the interval matrix. We propose a generalized Newton method for solving the linear complementarity problem with the regular interval matrix based on the nonlinear penalized equation. Further, we prove that this method is convergent. Numerical experiments are presented to show that the generalized Newton method is effective.
基金supported by the National Natural Science Foundation of China(Grant No.11602012)the 111 Project(Grant No.B07009)+1 种基金the Defense Industrial Technology Development Program(Grant No.JCKY2016601B001)and the China Postdoctoral Science Foundation(Grant No.2016M591038)
文摘A Newton iteration-based interval uncertainty analysis method(NI-IUAM) is proposed to analyze the propagating effect of interval uncertainty in multidisciplinary systems. NI-IUAM decomposes one multidisciplinary system into single disciplines and utilizes a Newton iteration equation to obtain the upper and lower bounds of coupled state variables at each iterative step.NI-IUAM only needs to determine the bounds of uncertain parameters and does not require specific distribution formats. In this way, NI-IUAM may greatly reduce the necessity for raw data. In addition, NI-IUAM can accelerate the convergence process as a result of the super-linear convergence of Newton iteration. The applicability of the proposed method is discussed, in particular that solutions obtained in each discipline must be compatible in multidisciplinary systems. The validity and efficiency of NI-IUAM is demonstrated by both numerical and engineering examples.
基金supported by the National Natural Science Foundation of China(No.51775230)。
文摘This paper presents a hybrid model reliability analysis method based on the damped Newton method with both random and interval variables to solve the hybrid structure reliability problem.The method combines an outer iterative solution and inner layer numerical calculation.In the outer iteration,the method seeks an optimized solution to the interval variable iterative by adding the boundary constraint condition based on the damped Newton optimization theory.In the inner layer solution,the method first reduces the dimension of the random variable through the dimension reduction method,then obtains the first four-order central moment of the function through the application of the Taylor expansion method,and finally calculates the reliability index of the structure according to the fourth-order moment calculation structure of the function.The results of a numerical example and an engineering ten-rod truss structure show that the proposed method can effectively solve the random-interval hybrid reliability problem and has better calculation accuracy than that of the two-layer iterative method.
