A threat assessment method for near-hypersonic weapons based on subjective and objective weights
基于主客观权重的临近高超声速武器威胁评估方法

摘要

临近空间高超声速武器具有飞行速度快、突防能力强、轨迹机动多变等突出特征，对传统防空反导体系构成了前所未有的严峻挑战。对其威胁程度的准确评估是实施合理资源分配与火力拦截的重要前提。针对高超声速武器威胁评估中单一赋权方法存在的固有不足，本文提出了一种融合模糊层次分析法与主成分分析法的主客观综合赋权威胁评估方法。首先，依据高超声速武器的飞行特性与物理属性，从多个维度建立威胁评估指标体系；其次，借助模糊层次分析法从专家经验中提取主观权重，同时运用主成分分析法从数据自身的方差特征中挖掘客观权重；最后，以离差最大化方法构建目标函数，并利用拉格朗日乘子法求解得到最优的组合权重系数。仿真实验中采用10批次高超声速目标仿真数据对所提方法进行验证，结果表明该主客观组合赋权模型能够更加客观、准确地反映目标的真实威胁排序，有效削弱了评估过程中的主观随意性与数据片面性对结果的影响。

Abstract

Near-space hypersonic weapons possess prominent features such as high flight speed, strong penetration capability, and highly variable maneuvering trajectories, posing unprecedented severe challenges to traditional air and missile defense systems. Accurate assessment of their threat level serves as an essential prerequisite for rational resource allocation and fire interception. To address the limitations of single weighting methods in hypersonic weapon threat assessment, we proposes a subjective-objective comprehensive weighting threat assessment method that integrates the fuzzy analytic hierarchy and principal component analysis. Firstly, we establish a threat assessment system based on the flight and physical attributes of hypersonic weapons. Second, subjective weights are learned from expert experience via the fuzzy analytic hierarchy process. Meanwhile, objective weights are computed from the data variance through principal component analysis. Finally, we construct an objective function by using the maximum deviation method, and the optimal combination weight coefficients are obtained by solving the function with the Lagrange multiplier method. To validate the effectiveness of proposed method, simulation experiments are conducted by using 10 batches of hypersonic target simulation data. The results demonstrate that the subjective-objective combination weighting model can reflect the true threat ranking of targets more objectively and accurately, effectively mitigating the influence of subjective arbitrariness and data one-sidedness on the assessment results.