遗传算法与蚂蚁算法的融合在神经网络中的研究

Research on the Combination of Genetic Algorithm and Ant Algorithm of Neural Networks

为了克服和改进传统BP算法的不足,发挥神经网络、遗传算法和蚂蚁算法各自的优势,本文提出了一种遗传算法和蚂蚁算法的融合在神经网络中二次训练的方法,并将融合算法应用于神经网络的权值训练中,采用遗传算法生成信息素分布,同时利用蚂蚁算法求精确解,并用神经网络二次训练得到最终结果,优势互补,获得了一种优化性能与时间性能共赢的有效算法。

To overcome and improve on the lack of the traditional BP algorithm, and take the advantage of ar- tificial neural network, genetic algorithm and ant algorithm, this paper puts forward a kind of second training method of the artificial neural network. By employing the combination of genetic algorithm and ant algorithm,it bring genetic algorithm into the learn weight training artificial neural network. It adopts genetic algorithm to give information pheromone to distribute and makes use of the ant algorithm to give the precision of the solution. Since the advantage of the two algorithms is complementary, it is more efficacious pheromone.