摘要： 在方案优化过程中，为消除随机因素的影响，需要进行蒙特卡洛仿真。为此，提出将置信区间引入到体系对抗仿真中，根据仿真过程中产生的数据样本确定蒙特卡洛仿真是否结束。依据数据样本的特性，对置信区间法进行相应的扩展，在获取满足精度需求的解的条件下，尽量缩减仿真实验次数。仿真实验结果表明，扩展置信区间法能够具备对蒙特卡洛仿真的自适应控制能力。
关键词:
体系对抗仿真,
专家经验法,
简单图线法,
置信区间法,
自适应控制
Abstract: In the process of solution optimization, in order to eliminate effect of random factors, Monte Carlo simulation is needed. This paper describes how to introduce confidence interval into system-of-systems combat simulation, so whether to end the Monte Carlo simulation based on data generated from simulation process can be determined. And based on the character of data, the confidence interval method is extended, so the number of experiments at the condition of meeting the precision requirement of solutions can be minimized. Simulation results show that the extended confidence interval method can control Monte Carlo simulation adaptively.
Key words:
system-of-systems combat simulation,
rule of thumb,
simple graphical method,
confidence interval method,
adaptive control
中图分类号:
吴红, 王磊, 杨峰, 王维平. 体系对抗仿真中实验次数的自适应控制方法[J]. 计算机工程, 2011, 37(3): 254-256.
TUN Gong, WANG Lei, YANG Feng, WANG Wei-Beng. Adaptive Control Method of Experiments’ Number in System-of-Systems Combat Simulation[J]. Computer Engineering, 2011, 37(3): 254-256.