Abstract: Training an artificial neural network is an optimization task since it is desired to find optimal weight set of a neural network in training process.Traditional training algorithms have some drawbacks such as getting in local minima and computational complexity.This paper introduces a kind use for training artificial feed-forward neural network based on Bee Colony Algorithm.Bee Colony Optimization algorithm is a simple，robust and population based stochastic optimization algorithm.The algorithm combines the exploration and exploitation processes effectively，and adopts a certain search strategy to skip from local optimization.The algorithm is successfully applied to XOR，N-Bit Parity and Encoder-Decoder problems，compared with the BP algorithm.Simulation results show that the proposed algorithm has better performance than the traditional GD algorithm and the LM algorithm.

Key words: bee colony optimization, training algorithm, feed-forward neural network

摘要： 训练人工神经网络的目的是调整各层的权重系数以达到最优，因而训练过程的实质是一项优化任务。传统的训练算法存在着容易陷入局部最优、计算复杂等缺陷。介绍一种训练前馈神经网络的蜜蜂群优化算法，它是一种简单、鲁棒性强的群体智能随机优化算法。该算法把探查和开发过程有效地结合在一起，并采取了跳出局部最优的搜索策略。成功地把该算法应用于神经网络训练的基本问题：异或问题、N位奇偶校验和编码解码问题，并与传统的BP算法进行比较。仿真实验证明其性能较传统的GD算法和LM算法更为优越。

关键词: 蜜蜂群优化, 训练算法, 前馈神经网络