基于人工蜂群算法的BP神经网络研究

BP Neural Network Based on Artificial Bee Colony(ABC) Algorithm

BP神经网络具有实现非线性映射特点和较强的容错能力、泛化能力等优点。然而,因为其采用了最速下降梯度寻优算法,在实际应用中往往出现收敛速度缓慢、时常陷入局部极小值等缺陷。人工蜂群算法是受蜂群个体间通过相互协作对既定目标进行寻优的群体行为启发提出的一种新型群智能优化算法,具有很好的全局收敛特性,其次有较强的自适应性、协作性、鲁棒性、快速性等特点。文中探讨用人工蜂群算法来优化BP神经网络算法,进一步提高BP神经网络性能。

BP neural network has the characteristics of nonlinear mapping and strong fault tolerance ability, generalization ability and so on. However, because it uses the steepest descent gradient optimization algorithm, the convergence rate is often slow, often caught in local minima and other defects in practical applications. Artificial bee colony algorithm is a novel swarm intelligence optimization algorithm, through mutual cooperation between individuals by the colony proposes a group behavior optimization for the target, with good convergence properties, strong adaptability, collaboration and fast etc. In this paper, the artificial bee colony algorithm is used to optimize the BP neural network algorithm, which can further improve the performance of BP neural network.