基于多分辨率仿真的多状态系统任务可靠性及成功性评估方法  

An assessment approach for multi-state system mission reliability and success probability based on multi-resolution simulation

作　　者：徐一帆 杨皓洁 吕建伟[1] 狄鹏[1] XU Yifan;YANG Haojie;LüJianwei;DI Peng(Department of Management Engineering and Equipment Economics,Naval University of Engineering,Wuhan 430033,China)

机构地区：[1]海军工程大学管理工程与装备经济系,武汉430033

出　　处：《系统工程理论与实践》2021年第5期1328-1342,共15页Systems Engineering-Theory & Practice

基　　金：国家自然科学基金(71401171);中国博士后科学基金(2019M653925)。

摘　　要：研究了装备系统研制早期在不同系统层次具有不完整、不确定信息条件下,系统任务可靠性及成功性的预估问题.通过逐层触发随机事件的分辨率解聚机制和重要性抽样的样本统计值修正方法,实现不同系统层次多分辨率数据的融合仿真.通过多分辨率条件下的受迫转移、失效偏倚等加速抽样方法,提高小概率事件的抽样效率,并通过修正抽样统计值,确保估计的无偏性.结合舰船系统航渡任务的算例研究,验证多分辨率仿真方法的有效性.发现该方法对于短任务时间或长任务时间,仿真效率均有不同程度的提高.基于仿真的灵敏性分析,评估了故障率、修复率及维修时间分布类型对系统可用度及任务成功性概率的影响程度.The problem of system mission reliability and success estimation under the condition of incomplete and uncertain information at different system levels in the early stage of weapon system development is studied.Through the resolution disaggregation mechanism of triggering random events layer by layer and the statistical value correction method of importance sampling,the fusion simulation of multi-resolution data at different system levels is realized.Additionally,through the accelerated sampling method such as forced transition and fault biasing under multi-resolution conditions,the sampling efficiency of small probability events is improved,and the sampling statistics are revised to ensure the unbiased estimation.By a case study of the ship system’s navigation mission,the effectiveness of the multi-resolution simulation method is verified.It is found that the presented method improved simulation efficiency to different degrees for short task time and long task time.Based on the simulation sensitivity analysis,the influence of failure rate,repair rate and the type of maintenance time distribution on the system availability and mission success probability were evaluated.

关 键 词：任务可靠性 任务成功性 多分辨率仿真 蒙特卡罗仿真 多状态系统 

分 类 号：O242.1[理学—计算数学]