No label defined (Q4510329)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 1521490
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | No label defined |
scientific article; zbMATH DE number 1521490 |
Statements
25 October 2000
0 references
initial value problem
0 references
spline collocation methods
0 references
numerical algorithm
0 references
stiff initial problems
0 references
0.8183851838111877
0 references
0.8063357472419739
0 references
0.7746714353561401
0 references
The initial value problem \[ y'(t)= f(t,y(t)),\quad t\in [0,T] ,\qquad y(0)=y_0, \tag{1} \] is considered, with \(f\in C^{p+1}([0,T],{\mathbb R}^m)\), \(m\geq 1\) and \(p\in {\mathbb N}\) sufficiently large. On the interval \([0,T]\) a uniform mesh \(\Delta: 0=t_0<t_1< \ldots<t_N=T\), \(h:=t_{k+1}=T_k\), \(h:=t_{k+1}-t_k\), \(k=0,\ldots,N-1\) is defined. Consider the collocation parameters \(c_1,c_2,\ldots,c_s\), \(s\geq 1\) in \([0,1]\) with the corresponding multiplicities \(\delta_1,\delta_2,\ldots,\delta_s\), strictly positive natural numbers. The higher derivatives of the right hand side in \((1)\) are: \[ f^{(0)}:=f, \qquad f^{(i+1)}:=f_t^{(i)}+f_y^{(i)}\cdot f,\quad i=0,{\l}~dots,p-1, \] where \(f_t^{(I)}\) is the partial derivative with respect to \(t\) and \(f_y^{(i)}\) is the Jacobian matrix with respect to the second vector variable \(y\). The deficient spline approximating the solution of \((1)\) is now defined by the conditions: \(\sigma_k: [t_k,t_{k+1}]\rightarrow {\mathbb R}^m\) \[ \sigma_k(t_k)= \sigma_{k-1}(t_k),\qquad \sigma_{k}^{(j)}(t_k+c_ih)= f^{j-1}(t_k+c_ih,\sigma_k(t_k+c_ic)), \tag{2} \] for \(k=0,\ldots,N-1\), \(i=1,\ldots,s\), \(j=1,\ldots, \delta_i\) and the starting conditions \(\sigma_{-1}^{(j)}=y^{(j)}(0)\) for \(j=0,\ldots,p+1\). The global deficient spline approximating the solution of \((1)\) is now recovered by \(\sigma: [0,t]\rightarrow {\mathbb R}^m\) by \(\sigma|_{[t_k,t_{k+1}]}:=\sigma_k, \) for \(k=0,\ldots,N-1\). The above described method is called \(s\)-stage multiderivatve spline collocation (MSC) method. The multiderivative spline collocation methods are very accurate instruments for numerically solving differential equations. Due to their implicit character, their implementation is computationally very costly, especially for nonlinear initial value problems. For this reason, they must perform well for problems on which the ''cheap'' explicit methods fail, i.e.\ for stiff initial problems. In this note, some necessary conditions for these properties are given.
0 references