基于改进萤火虫优化算法的BP神经网络目标群威胁判断

以舰艇防空作战目标选择决策和规划需求为背景，针对萤火虫算法求解精度不高且收敛速度较慢的问题，提出可动态调整步长的改进萤火虫优化算法。在改进萤火虫优化算法的基础上，建立基于改进萤火虫优化算法的BP神经网络目标群威胁判断结构模型。通过改进萤火虫算法优化BP神经网络的初始权值和阈值，能够更好地预测测试集。实验结果表明，该方法可快速、准确地实现目标群威胁判断。

Threat Assessment of Target Group Using Improved Glowworm Swarm Optimization and BP Neural Network

Setting the ship air defense system as a background, aiming at the problem of the accuracy is not high and the convergence is slow in glowworm swarm optimization. The glowworm swarm optimization can adjust the adaptive step size dynamically. Establishing judge model improved the glowworm swarm optimization and BP neural network based on the improved glowworm swarm optimization algorithm. Optimization of BP neural network by improving the firefly algorithm the initial weights and thresholds, prediction can be better on the test set. Experimental results show that, the method can realize the threat assessment of target group quickly and accurately.