分布式检测系统的混合优化设计方法

A hybrid optimization method for distributed detection

提出了一种用于Ｎｅｙｍａｎ－Ｐｅａｒｓｏｎ准则下分布式检测系统优化设计的混合优化方法．该方法同时使用进化计算和Ｐｏｗｅｌ方向加速法进行搜索，无须使用目标函数的导数信息，适用于非线性优化问题以及导数信息难于计算的优化问题，具有适用性好、精度高的优点．

A hybrid method is proposed for the optimization of the distributed detection system under Neyman Pearson criteria. It optimizes the object function by using evolutionary computation and Powell method, does not need the derivation information of the object function, and can be applied to nonlinear optimization problems or problems whose derivation information is hard to calculate.