Abstract: Basic Firefly Algorithm（FA）has some bugs such that it is easy to fall into local optimum and the slow convergence speed. In order to overcome these shortcomings, the paper puts forward a Firefly Algorithm with Improved Evolution Mechanism（IEMFA）. The proposed algorithm is used to optimize the weight value of BP neural network and the result shows that the algorithm expedites the convergence rate, improves the precision and has a better global searching ability.

Key words: improved evolutionary mechanism, error Back Propagation（BP） neural network, firefly algorithm

摘要： 基本萤火虫算法存在容易陷入局部最优及收敛速度低的问题，提出了一种改进进化机制的萤火虫算法（IEMFA）。在群体进化过程中赋予萤火虫改进的位置移动策略，并利用改进后的萤火虫算法来优化传统BP神经网络的网络参数。测试结果表明，基于改进萤火虫算法的BP神经网络具有更好的收敛速度和精度。

关键词: 进化机制, 误差反向传播（BP）神经网络, 萤火虫算法