f******n
发帖数: 176 | 1 Any comments about this package? Thank you! |
|
f**l
发帖数: 2041 | 2 For stiff problem, I suggest you try DASSL. According to a recent
review, it is the most efficient implicit solver. If your problem
is large scale, you might want to use DASPK, which uses Krylov method
to solve linear systems. But I heard that the preconditioner could
be a hard problem when using DASPK.
For nonstiff problem, Adams method is good enough. There are many
software using that. I guess Shampine's software should be good, since
she wrote all the ode solvers for matlab. |
|
f**l
发帖数: 2041 | 3 为什么要关注这个定义呢? 关于stiff的定义可以有很多种. 但关键还是在于
数值稳定性的考虑. 你所说的都是把方程线性化之后的分析. 这些分析只是为了
给人一个直觉的基础罢了. 数学要是只关心定义, 就失去其应用的一面了.
对你的问题, 你直接用ode15s就是了. 一般用matlab来解的都不会太复杂.
ode15s是建立在BDF公式上的变阶变步长算法, 由C. W. Gear早先提出的.
如果问题比较大, 建议使用CVODE, LSODE, DASSL, DASPK等成熟的软件. 严肃的
计算一般是不用matlab搞的.
消除stiffness取决于问题. 如果你知道是什么部分导致一些变量很快地进入
动态平衡, 把这部分用代数的方法取代, 就是了. 这个具体到问题可能很复杂.
the
some |
|
h****J
发帖数: 56 | 4 My fault. I had typos earlier.
Using MatLab to do ODE simulation is very simple; I don’t think you need
any extra software kits. Never used MatLab for parameter estimation before,
but had some experience with FORTRAN. There is a software solver called
LSODE, see below for details. I believe this is a more professional package
but it could be tricky to use. You may need someone to help you out, if you
didn’t have much engineering background.
I guess you are doing some compartment PK modeling?
=== |
|
h****J
发帖数: 56 | 5 Forgot to mention: if you want to use LSODE, they can be downloaded from
Lawrence Livermore National Laboratory, and they are free.
http://www.llnl.gov/CASC/software.html.
,
package
you |
|
