Structure-preserving algorithms for a class of dynamical systems

被引:1
|
作者
Wang, Ling-shu [1 ]
Feng, Guang-hui [2 ]
机构
[1] Hebei Univ Econ & Business, Sch Math & Stat, Shijiazhuang 050061, Peoples R China
[2] Mech Engn Coll, Shijiazhuang 050003, Peoples R China
来源
关键词
SRK methods; numerical experiment; backward error analysis;
D O I
10.1007/s10255-006-0361-0
中图分类号
O29 [应用数学];
学科分类号
070104 ;
摘要
In this paper, we study structure-preserving algorithms for dynamical systems defined by ordinary differential equations in R-n. The equations are assumed to be of the form y = A(y) + D(y) + R(y), where A(y) is the conservative part subject to < A(y), y > = 0; D(y) is the damping part or the part describing the coexistence of damping and expanding; R(y) reflects strange phenomenon of the system. It is shown that the numerical solutions generated by the symplectic Runge-Kutta(SRK) methods with b(i) > 0 ( i = 1, ..., s) have long-time approximations to the exact ones, and these methods can describe the structural properties of the quadratic energy for these systems. Some numerical experiments and backward error analysis also show that these methods are better than other methods including the general algebraically stable Runge-Kutta(RK) methods.
引用
收藏
页码:161 / 176
页数:16
相关论文
共 50 条