摘要
针对空战中机动决策速度慢、准确性低问题,提出基于强化麻雀搜索神经网络的空战机动决策方法,该方法分别考虑角度、距离、高度等因素构造相应的态势函数,将几种态势函数结合起来并加权得到态势评估函数,利用神经网络的黑盒部分计算各个态势函数的权值,利用混沌初始化和小孔成像反向学习策略强化麻雀搜索算法,再利用其特性去优化神经网络,将麻雀的适应度函数与神经网络的权值和阈值建立直接的映射关系,从而获得准确的态势评估函数,再将得到的态势评估函数结合博弈论模型得到博弈态势值,并由此来进行空战机动决策。仿真表明,该方法在与粒子群和遗传算法优化的神经网络模型相比之下,决策速度更快、准确性高,从而能获得空战优势,以获得最终的空战胜利。
For air combat maneuvering decision making in the low speed,low accuracy problem,air combat maneuvering decision-making method based on intensified sparrow search neutral network is put forward and the method,respectively,considers such factors as tectonic Angle,distance,height to construct the corresponding situation function,several situation functions are combined and weighted to get the function of situation assessment,and then black box parts of the neutral network is used to calculate weights of each situation function.Chaos initialization and small hole imaging reverse learning strategies are used to strengthen the sparrow search algorithm,the sparrow fitness function and neural network weights and threshold are used to establish the direct mapping relationship,so as to obtain accurate situation assessment function,then the obtained situation assessment function is combined with game theory model to get the game situation values,and then the air combat maneuver decision-making is made.Simulation results show that compared with the neural network model optimized by particle swarm optimization and genetic algorithm,the method adopted in this paper has faster decision-making speed and higher accuracy,so as to obtain the superiority in air combat and to obtain the final victory in air combat.
作者
刘庆利
乔晨昊
杨国强
张振亚
LIU Qingli;QIAO Chenhao;YANG Guoqiang;ZHANG Zhenya(Key Laboratory of Communication and Network,Dalian University,Dalian 116000,China)
出处
《火力与指挥控制》
CSCD
北大核心
2023年第1期84-91,98,共9页
Fire Control & Command Control
关键词
空战
神经网络
强化麻雀搜索算法
机动决策
反向学习策略
air combat
neural network
intensified sparrow search algorithm
maneuvering decisionmaking
reverse learning strategies
