用仿真方法对随机系统优化设计

对一个随机系统(或过程)进行优化设计的难点在于随机变量的处理和计算，另外，由于待优化的变量往往不直接包含在目标函数中，优化算法选用受到一定限制。为解决此问题可采用计算机仿真方法模拟实际的动态随机过程，多次运行此过程可获得足够的样本，可计算出统计量。用遗传算法控制待优化的变量，可以较快的获得满意的随机系统的优化设计。

Optimization Design for Stochastic System by Simulation Method

The difficulty of optimization design for stochastic system (or process) is the processing and calculation to stochastic variable. In addition, because the preparative optimization variable isn't always directly included in aim function, the choosing of optimization algorithms is restricted. It can be resolved by simulation method of imitating actual dynamic stochastic systems. Operating the system for many times can acquire enough samples and then figure out statistic value. That genetic algorithms control preparative optimization variable can quickly win satisfactory optimization design of stochastic system.