欢迎访问《兵工学报》官方网站,今天是 分享到:

兵工学报 ›› 2016, Vol. 37 ›› Issue (4): 727-734.doi: 10.3969/j.issn.1000-1093.2016.04.022

• 论文 • 上一篇    下一篇

基于多智能体遗传算法优化的航空电子设备状态组合预测

赵建忠1, 欧阳中辉1, 张磊2, 赵建印1   

  1. (1.海军航空工程学院 兵器科学与技术系, 山东 烟台 264001; 2.海军航空工程学院 科研部, 山东 烟台 264001)
  • 收稿日期:2015-07-03 修回日期:2015-07-03 上线日期:2016-06-20
  • 通讯作者: 赵建忠 E-mail:zjznavy@163.com
  • 作者简介:赵建忠(1978—),男,讲师
  • 基金资助:
    总装备部基础科研项目(2014年)

Combined Prediction on Avionics State Optimized by MAGA

ZHAO Jian-zhong1, OUYANG Zhong-hui1, ZHANG Lei2,ZHAO Jian-yin1   

  1. (1.Department of Ordnance Science and Technology,Naval Aeronautical and Astronautical University, Yantai 264001, Shandong, China;2.Department of Scientific Research, Naval Aeronautical and Astronautical University, Yantai 264001, Shandong, China)
  • Received:2015-07-03 Revised:2015-07-03 Online:2016-06-20
  • Contact: ZHAO Jian-zhong E-mail:zjznavy@163.com

摘要: 针对传统单一预测方法预测航空电子设备状态的不足,提出了将隐马尔可夫模型(HMM)和最小二乘支持向量机(LS-SVM)相结合的组合预测方法。采用多智能体遗传算法(MAGA)对HMM参数进行训练优化,克服了Baum-Welch算法易陷入局部最优解的缺陷,并在HMM建模过程中引入状态条件概率,以降低不确定性因素造成的影响。采用MAGA估计LS-SVM模型参数,并在参数估计的过程中采用剪枝法实现LS-SVM的稀疏性,从而达到提高LS-SVM泛化性能的目的。在此基础上构建了航空电子设备状态组合预测模型。实例分析结果验证了组合预测模型在预测精度、计算速度和稳定性方面的优势。

关键词: 飞行器仪表、设备, 参数估计, 隐马尔可夫模型, 最小二乘支持向量机, 多智能体遗传算法, 状态预测

Abstract: A combined prediction method based on hidden Markov model (HMM) and least square support vector machine (LS-SVM) is presented for prediction of avionics states. Multi-agent genetic algorithm (MAGA) is used to estimate HMM parameters to overcome the problem of that Baum-Welch algorithm is easy to fall into local optimal solution. The conditional probability of states is introduced into the HMM modeling to reduce the effect of uncertainty factor. MAGA is used to estimate the parameters of LS-SVM model, and the pruning algorithm is used for achieving the sparse approximation of LS-SVM in parameter estimation, thus achieving the objective of improving the generalization performance of LS-SVM. On this basis, a combined prediction model of avionics state is established. The analysis results show the combined prediction model has high prediction accuracy, calculating speed and stability.

Key words: instrument and equipment of aerocraft, parameter estimation, hidden Markov model, least square support vector machine, multi-agent genetic algorithm, state prediction

中图分类号: