基于多Markov链的效果评估模型研究  

Study of Multi-Markov Chains to Estimating Effect of Naval Mine-sweeping

作　　者：李华[1] 龚沈光[1] 李庆民[1] 刘君[1] 

机构地区：[1]海军工程大学兵器工程系,武汉430033

出　　处：《武汉理工大学学报（交通科学与工程版）》2007年第4期699-702,共4页Journal of Wuhan University of Technology（Transportation Science & Engineering）

基　　金：国防预研项目资助

摘　　要：基于单Markov链的扫雷作战效果评估模型是一个理论意义大于实际的模型,离实际的多Markov扫雷过程还有相当的距离.文中从正交分解的角度,指出Markov评估模型只能进行定性评估.提出了用于定性评估的多Markov模型,从"分解基"相关性的角度提出了多链模型结构选择的方式,实现了基于遗传算法的模型解法,计算结果表明:在引入先验知识后,能在小样本数据条件下取得较为准确的评估结果.At present,the evaluating model of mine-sweeping campaign effect based on single Martkov chain is a theoretic model and short of practice. It still has a largish distance to the multi Markov mine sweeping process, which had been applied in practice. This text starts with the orthogonal basis and points out that Markov evaluating model only can get a determined evaluating result. It suggests the multi-Markov model for determining evaluating,the way how to choose the structure of multi-Markov model, and solution based genetic algorithm. The result indicats; after importing preknowledge, quite precise evaluating result can be got under the condition of small swatch data.

关 键 词：扫雷作战效果 MARKOV链 遗传算法 

分 类 号：TJ01[兵器科学与技术—兵器发射理论与技术]