一种基于多agent系统的生化系统分布式随机仿真算法  

A multi-agents-based distributed algorithm for stochastic simulation of biochemical reaction systems

作　　者：钮俊清[1] 郑浩然[1] 陈久生[1] 马猛[1] 王煦法[1] 

机构地区：[1]中国科学技术大学计算机科学与技术系,合肥230027

出　　处：《北京生物医学工程》2007年第5期485-488,共4页Beijing Biomedical Engineering

基　　金：中国科学院知识创新工程重要方向项目(KSCX2-SW-329)资助

摘　　要：系统生物学关注于对生化反应网络进行定量的数学建模和计算机仿真,这是理解生化网络的拓扑结构和动态行为的主要手段。Gillespie的随机仿真算法(SSA)是仿真均匀生化反应系统的一种标准的算法,而SSA算法最大的问题在于计算代价过高。基于并行思想提出一种基于多agent系统实现的分布式随机仿真算法(DSSA),利用分布式计算来提高SSA算法的计算性能,为建模与仿真大规模生化反应系统提供了一种有效的方式。实验显示DSSA算法在时间性能上带来显著的提升。Systems biology concerned with the quantitative mathematical modelling and computer simulation of biochemical reaction networks, which was the principal means of understanding biochemical network topology and dynamic behaviour. Gillespie＇ s stochastic simulation algorithms（SSA） was a standard algorithms to simulate well-stirred biochemical reaction system, however the biggest problem of SSA algorithm lied in the excessive cost of computation. Based on the parallel theory we proposed a new distributed-based stochastic simulation algorithms （DSSA） using multi-agents system and distributed computing to improve computing performance SSA, and DSSA was an effective way to model and simulate large biochemical reaction systems. Experiments showed DSSA algorithm was able to improve time performance significantly.

关 键 词：系统生物学 随机仿真算法 多agents分布式仿真 

分 类 号：R318.04[医药卫生—生物医学工程]