改进粒子群优化BP神经网络的目标威胁估计

Target threat assessment based on BP neural network optimized by modified particle swarm optimization

为了提高目标威胁估计精度,提出一种运用改进粒子群算法优化BP神经网络的方法。为了避免陷入局部极值,将变异过程引入粒子群算法中,并对相关参数进行优化,形成改进粒子群算法,对BP神经网络的初始权值和阈值进行优化。利用样本数量不同的训练集对网络进行训练,并用60组测试集数据对网络进行验证。实验结果表明,改进粒子群优化BP神经网络目标威胁估计算法具有更高的预测精度,在训练样本数量较小时能够获得较好的预测能力,可以有效地完成目标威胁估计。

An algorithm for target threat assessment based on Back Propagation （BP） neural network optimized by Modified Particle Swarm Optimization （MPSO ） is proposed to improve the prediction accuracy of target threat. In this MPSO algorithm, mutation operator and optimization for several parameters are introduced in PSO to avoid the particle plunging into the local optimization. The MPSO algorithm is employed to optimize the initial weights and thresholds of the BP neural network. Then the BP neural network optimized by MPSO is trained by training sets of different sample sizes. 60 sets of target threat data are adopted to test the performance of MPSO - BP in target threat prediction. Experimental results show that the prediction accuracy of target threat assessment algorithm based on MPSO-BP is higher than that based on some traditional algorithms, which proves the efficiency of the proposed algorithm in solving target threat assessment problem in spite of the small sample size of training set.