Time dependent solution for acceleration of tau-leaping
Time dependent solution for acceleration of tau-leaping
复制标题
DOI：
10.1016/j.jcp.2012.10.036
复制
发表时间：
2013-02-15
影响因子：
4.1
通讯作者：
Petzold, Linda R.
中科院分区：
文献类型：
--
作者：
Fu, Jin;Wu, Sheng;Petzold, Linda R.
τ跳跃法是加速化学反应体系离散随机模拟的有效方法。然而，当涉及快速反应时，这种方法的加速可能相当有限。解决这个问题的一种方法是应用随机准稳态假设。但是，我们在使用这个假设时必须小心。如果快速子系统不能足够快地达到稳定分布，则准稳态假设将误差传播到模拟中。为了避免这些错误，我们建议使用时间相关的解决方案，而不是准稳态。一般来说，对于任意网络，时间相关的解是不容易得到的。然而，对于一些常见的图案，我们确实有时间依赖的解决方案。我们推导出这些图案的时间依赖性解决方案，然后展示它们如何与tau跳跃一起使用，以实现大幅加速，包括血液凝固的现实模型。虽然该方法是复杂的，我们已经自动化。（C）2012 Elsevier Inc. All rights reserved.
The tau-leaping method is often effective for speeding up discrete stochastic simulation of chemically reacting systems. However, when fast reactions are involved, the speed-up for this method can be quite limited. One way to address this is to apply a stochastic quasi-steady state assumption. However we must be careful when using this assumption. If the fast subsystem cannot reach a steady distribution fast enough, the quasi-steady-state assumption will propagate error into the simulation. To avoid these errors, we propose to use the time dependent solution rather than the quasi-steady-state. Generally speaking, the time dependent solution is not easy to derive for an arbitrary network. However, for some common motifs we do have time dependent solutions. We derive the time dependent solutions for these motifs, and then show how they can be used with tau-leaping to achieve substantial speed-ups, including for a realistic model of blood coagulation. Although the method is complicated, we have automated it. (C) 2012 Elsevier Inc. All rights reserved.