In [1], a class of multiderivative block methods (MDBM) was studied for the numerical solutions of stiff ordinary differential equations. This paper is aimed at solving the problem proposed in [1] that what conditions...In [1], a class of multiderivative block methods (MDBM) was studied for the numerical solutions of stiff ordinary differential equations. This paper is aimed at solving the problem proposed in [1] that what conditions should be fulfilled for MDBMs in order to guarantee the A-stabilities. The explicit expressions of the polynomialsP(h) and Q(h) in the stability functions h(h)=P(h)/Q(h)are given. Furthermore, we prove P(-h)-Q(h). With the aid of symbolic computations and the expressions of diagonal Fade approximations, we obtained the biggest block size k of the A-stable MDBM for any given l (the order of the highest derivatives used in MDBM,l>1)展开更多
In recent years, the derivation of Runge-Kutta methods with higher derivatives has been on the increase. In this paper, we present a new class of three stage Runge-Kutta method with first and second derivatives. The c...In recent years, the derivation of Runge-Kutta methods with higher derivatives has been on the increase. In this paper, we present a new class of three stage Runge-Kutta method with first and second derivatives. The consistency and stability of the method is analyzed. Numerical examples with excellent results are shown to verify the accuracy of the proposed method compared with some existing methods.展开更多
Based on newly developed weight-based smoothness detectors and non-linear interpolations designed to capture discontinuities for the multiderivative com-bined dissipative compact scheme(MDCS),hybrid linear and nonline...Based on newly developed weight-based smoothness detectors and non-linear interpolations designed to capture discontinuities for the multiderivative com-bined dissipative compact scheme(MDCS),hybrid linear and nonlinear interpolations are proposed to form hybrid MDCS.These detectors are derived from the weights used for the nonlinear interpolations and can provide suitable switches between the linear and the nonlinear schemes to realize the characteristics for the hybrid MDCS of capturing discontinuities and maintaining high resolution in the region without large discontinuities.To save computational cost,the nonlinear scheme with characteris-tic decomposition is only applied in the detected discontinuities region by specially designed hybrid strategy.Typical tests show that the hybrid MDCS is capable of cap-turing discontinuities and maintaining high resolution power for the smooth region at the same time.With the satisfaction of the geometric conservative law(GCL),the MDCS is further applied on curvilinear mesh to present its promising capability of handling pragmatic simulations.展开更多
In this paper,the necessary and sufficient conditions for generalone-step m ethods to be exponentially fitted atq0∈C aregiven.A classofm ultiderivative hybrid one-step m ethods of order at leasts+ 1 is constructed ...In this paper,the necessary and sufficient conditions for generalone-step m ethods to be exponentially fitted atq0∈C aregiven.A classofm ultiderivative hybrid one-step m ethods of order at leasts+ 1 is constructed w ith s+ 1 param eters,w here sis the order of derivative.The necessary and sufficient conditions for these m ethods to be A-stable and exponentially fitted is proved.Furtherm ore,a class ofA-stable 2 param eters hybrid one-step m ethods oforderatleast 8 are constructed,w hich use 4th order derivative.These m ethods are exponentially fitted atq0 if and only if its fitted function f(q) satisfies f(q0)= 0.Finally,an A-stable exponentially fitted m ethod oforder 8 is obtained.展开更多
This paper continues to study the explicit two-stage fourth-order accurate time discretizations[5-7].By introducing variable weights,we propose a class of more general explicit one-step two-stage time discretizations,...This paper continues to study the explicit two-stage fourth-order accurate time discretizations[5-7].By introducing variable weights,we propose a class of more general explicit one-step two-stage time discretizations,which are different from the existing methods,e.g.the Euler methods,Runge-Kutta methods,and multistage multiderivative methods etc.We study the absolute stability,the stability interval,and the intersection between the imaginary axis and the absolute stability region.Our results show that our two-stage time discretizations can be fourth-order accurate conditionally,the absolute stability region of the proposed methods with some special choices of the variable weights can be larger than that of the classical explicit fourth-or fifth-order Runge-Kutta method,and the interval of absolute stability can be almost twice as much as the latter.Several numerical experiments are carried out to demonstrate the performance and accuracy as well as the stability of our proposed methods.展开更多
Let X be a Hilbert space with the real or complex base field K, the inner product 【·,·】 and the corresponding norm ‖·‖. For any given matrix A = [a<sub>ij</sub>] ∈K<sup>p×q&l...Let X be a Hilbert space with the real or complex base field K, the inner product 【·,·】 and the corresponding norm ‖·‖. For any given matrix A = [a<sub>ij</sub>] ∈K<sup>p×q</sup>, we can define a linear mapping （?）: X<sup>q</sup>→X<sup>p</sup>:展开更多
文摘In [1], a class of multiderivative block methods (MDBM) was studied for the numerical solutions of stiff ordinary differential equations. This paper is aimed at solving the problem proposed in [1] that what conditions should be fulfilled for MDBMs in order to guarantee the A-stabilities. The explicit expressions of the polynomialsP(h) and Q(h) in the stability functions h(h)=P(h)/Q(h)are given. Furthermore, we prove P(-h)-Q(h). With the aid of symbolic computations and the expressions of diagonal Fade approximations, we obtained the biggest block size k of the A-stable MDBM for any given l (the order of the highest derivatives used in MDBM,l>1)
文摘In recent years, the derivation of Runge-Kutta methods with higher derivatives has been on the increase. In this paper, we present a new class of three stage Runge-Kutta method with first and second derivatives. The consistency and stability of the method is analyzed. Numerical examples with excellent results are shown to verify the accuracy of the proposed method compared with some existing methods.
基金supported by the National Key Research and Development Plan(grant No.2016YFB0200700)the National Natural Science Foundation of China(grant Nos.11372342,11572342,and 11672321)the National Key Project GJXM92579.
文摘Based on newly developed weight-based smoothness detectors and non-linear interpolations designed to capture discontinuities for the multiderivative com-bined dissipative compact scheme(MDCS),hybrid linear and nonlinear interpolations are proposed to form hybrid MDCS.These detectors are derived from the weights used for the nonlinear interpolations and can provide suitable switches between the linear and the nonlinear schemes to realize the characteristics for the hybrid MDCS of capturing discontinuities and maintaining high resolution in the region without large discontinuities.To save computational cost,the nonlinear scheme with characteris-tic decomposition is only applied in the detected discontinuities region by specially designed hybrid strategy.Typical tests show that the hybrid MDCS is capable of cap-turing discontinuities and maintaining high resolution power for the smooth region at the same time.With the satisfaction of the geometric conservative law(GCL),the MDCS is further applied on curvilinear mesh to present its promising capability of handling pragmatic simulations.
文摘In this paper,the necessary and sufficient conditions for generalone-step m ethods to be exponentially fitted atq0∈C aregiven.A classofm ultiderivative hybrid one-step m ethods of order at leasts+ 1 is constructed w ith s+ 1 param eters,w here sis the order of derivative.The necessary and sufficient conditions for these m ethods to be A-stable and exponentially fitted is proved.Furtherm ore,a class ofA-stable 2 param eters hybrid one-step m ethods oforderatleast 8 are constructed,w hich use 4th order derivative.These m ethods are exponentially fitted atq0 if and only if its fitted function f(q) satisfies f(q0)= 0.Finally,an A-stable exponentially fitted m ethod oforder 8 is obtained.
基金partially supported by the Special Project on Highperformance Computing under the National Key R&D Program(No.2020YFA0712002)the National Natural Science Foundation of China(No.12126302,12171227).
文摘This paper continues to study the explicit two-stage fourth-order accurate time discretizations[5-7].By introducing variable weights,we propose a class of more general explicit one-step two-stage time discretizations,which are different from the existing methods,e.g.the Euler methods,Runge-Kutta methods,and multistage multiderivative methods etc.We study the absolute stability,the stability interval,and the intersection between the imaginary axis and the absolute stability region.Our results show that our two-stage time discretizations can be fourth-order accurate conditionally,the absolute stability region of the proposed methods with some special choices of the variable weights can be larger than that of the classical explicit fourth-or fifth-order Runge-Kutta method,and the interval of absolute stability can be almost twice as much as the latter.Several numerical experiments are carried out to demonstrate the performance and accuracy as well as the stability of our proposed methods.
基金Project supported by the National Natural Science Foundation of China
文摘Let X be a Hilbert space with the real or complex base field K, the inner product 【·,·】 and the corresponding norm ‖·‖. For any given matrix A = [a<sub>ij</sub>] ∈K<sup>p×q</sup>, we can define a linear mapping （?）: X<sup>q</sup>→X<sup>p</sup>:
