基于改进模糊神经推理的武器装备系统组合决策

Portfolio decision of weapon system based on improved fuzzy neural inference

针对武器装备系统组合决策问题,提出了一种新的基于自适应模糊神经推理的组合决策模型。首先用由高斯函数表示的模糊集定量描述武器系统的作战效能和敏捷性;接着基于波士顿投资组合矩阵进行武器系统模糊分类,建立自适应神经网络;最后利用鸟群算法优化模型相关参数。在样本数据库上的仿真结果表明,该方法可以反映武器系统组合状态,使决策者可以根据需求对组合策略进行调整更新。此外鸟群算法优化后的模型能够在一定程度上提高分类精度,与传统模型相比,具有更低的均方误差和更高的误差容忍率。

To solve the problem of weapon system portfolio decision-making,this paper proposed a new portfolio decision mo-del based on adaptive fuzzy neural inference.First,it described combat effectiveness and agility of the system by fuzzy sets represented by Gaussian functions.Then,it classified the weapon systems and the adaptive fuzzy neural model established based on Boston’s portfolio matrix.Finally,it optimized the parameters which were related to the model by Satin bowerbird algorithm.The simulation result on the sample database shows that the method can reflect the state of the system of systems,so that the decision maker can adjust and update the portfolio scheme according to the demand.In addition,the optimized model based on Satin bowerbird algorithm can improve the classification accuracy to some extent,and has lower mean square error and higher error tolerance than the traditional model.