基于类型2区间模糊 K 近邻分类器的动态武器目标分配方法研究

摘要

Dynamic weapon-target assignment (WTA)problem is one of the critical problem when process-ing the battlefield command control and decision.The WTA algorithm makes the decision at the gap of two at-tack periods,which needs to be calculated efficiently.When using the battlefield assistant system to make WTA decision,it is a reasonable concept to utilize the machine learning method,make new decision from the known decision to avoid the need of searching new target assignment all over again.By using this concept,an interval type-2 fuzzy K-nearest neighbors(IT2FKNN)classifier usage on the WTA problem is proposed,by using the result from branch and bound to train the model,a parrallel IT2FKNN classifier is built to infer the assignment result.The proposed method is tested to be able to make a quick decision on the WTA problem.%动态武器目标分配问题是战场指挥控制决策中的关键问题。由于动态武器目标分配算法是在攻击间隙所做的决策，对计算时间的实时性要求较高。解决这一问题，可以采用机器学习的方法基于战场辅助决策系统的武器目标分配，从已知的决策中推理生成出新的决策，而不必每个步骤中都重新搜索新的目标分配方案。根据这种思路，提出了一种基于类型2区间模糊 K 近邻分类器的武器目标分配方法，利用分支定界法得到的分配方案作为训练样本，通过构造并行运行的类型2区间模糊 K 近邻分类器来推导目标分配结论，实现了快速决策的目的。