不确定环境下潜艇综合防御鱼雷的作战决策分析与优化方法  

Operational Decision Analysis and Optimization Method of Submarine Integrated Defense against Torpedo in Uncertain Environment

作　　者：杨静 陆铭华[2] 郭力强[2] 马洁琼 吴金平 张会[2] YANG Jing;LU Minghua;GUO Liqiang;MA Jieqiong;WU Jinping;ZHANG Hui(College of System Engineering,National University of Defense and Technology,Changsha 410073,Hunan,China;Navy Submarine College,Qingdao 266041,Shandong,China;Unit 92020 of PLA,Qingdao 266001,Shandong,China)

机构地区：[1]国防科技大学系统工程学院,湖南长沙410073 [2]海军潜艇学院,山东青岛266041 [3]92020部队,山东青岛266001

出　　处：《兵工学报》2024年第2期564-573,共10页Acta Armamentarii

摘　　要：针对潜艇综合防御决策时存在的目标类型不明、目标运动信息解算不明、对抗环境信息不明3类信息不确定问题,构建基于不确定信息的潜艇防御鱼雷组合优化框架,设计基于蒙特卡洛模拟的随机搜索优化模型,采用基于离散边界收缩的粒子群优化(Particle Swarm Optimization,PSO)方法计算局部最优决策。建立战术对抗仿真模型对潜艇防御鱼雷策略进行选择与优化,针对不同鱼雷报警位置和行动策略,以安全余量作为评价标准,综合考虑3类不明条件下使用潜艇转向+定深规避与潜艇转向+使用水声器材防御两种策略,并结合PSO方法与过程仿真分别给出两类策略的选择边界条件及不同作战态势下的决策建议。防御效能分析结果表明:目标解算不收敛是最主要的防御行动影响因素,基于PSO方法的潜艇转向角度决策精度为±10°/20°;在相同的条件下,组合优化框架的决策加速比相比于传统搜索方法达到92~156,相比于并行计算方法达到10~16,在满足决策实时性的基础上可以为潜艇防御鱼雷争取更大的防御决策时间。There are three kinds of uncertain information,including unknown target,unknown target motion and unknown underwater environment,in submarine defense situation.A combined optimization framework based on uncertain information for making decisions of submarine defense against torpedoes is constructed.A random search optimization model based on Monte Carlo methods is designed,and then the particle swarm optimization(PSO)method based on discrete boundary contraction is used to calculate the local optimal decision instead of the best resolution.A tactical confrontation simulation model is established to select and optimize the submarine defense torpedo strategies.In details,for different torpedo alarm positions and action strategies,the two defense strategies,namely steering-diving and steering-bait under three types of uncertainty conditions,are comprehensively considered by taking“safety margin”as an evaluation standard.The selection boundaries of the two strategies and the decision-making suggestions under different combat situations are given,respectively.The experiment demonstrates that,the non-convergence of target solution is the most important factor affecting the defense action,and the decision-making accuracy of submarine steering angle based on PSO is±10°/20°;uder the same conditions,the acceleration ratio of the proposed framework can reach 92-156 compared with that of the traditional search method,and 10-16 compared with that of the parallel computing method.The proposed method could buy more time for decision making under the real-time requirements.

关 键 词：不确定信息 潜艇防御鱼雷 粒子群优化方法 过程仿真 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]