摘要
针对现代空战效能评估多层次、多因素、非线性耦合的复杂特点,首先引入空战效能评估完整、合理的指标体系,然后建立基于人工神经网络(ANN)的空战效能评估模型。一方面利用ANN极强的函数逼近能力和泛化能力来映射空战效能指标与效能值之间复杂的非线性关系。同时利用ANN的容错能力来增强处理在评估过程中由于指标信息含糊、样本不完整或指标间非线性耦合而导致的不确定性误差的能力。使其具有较好的推广和预测能力。最后进行的软件仿真计算表明,该方法能够实现高效、可靠的空战效能评估。
Considering the multi-level, non-linear and coupling of the complex characteristics of modem air combat effectiveness evaluation, we adopted a complete and reasonable index system for air combat effectiveness and built up an effectiveness evaluation model based on Artificial Neural Network （ANN）. The high approximation and generalization ability of ANN function was used to map the complex, nonlinear relations between each effectiveness index and the effectiveness value, while the fault-tolerant ability of ANN was used to enhance the capability to deal with the uncertain errors caused by ambiguous index information, incomplete samples or nonlinear coupling between different indexes. Therefore the method has fine predicting capability. Simulation result showed that the method can implement highly efficient and reliable effectiveness evaluation of the air combat.
出处
《电光与控制》
北大核心
2010年第4期26-29,共4页
Electronics Optics & Control
基金
"十一五"预研基金项目(KJ-050402011)
空军预研项目(402050301)
关键词
空战效能
效能评估
人工神经网络
BP网络
air combat effectiveness
effectiveness evaluation
Artificial Neural Network （ANN）
BP neural network
