基于蚁群优化算法的BP神经网络的RPROP混合算法仿真的研究  

Research on RPROP Hybrid Algorithm Model of BP Neural Network Based on Ant Colony Optimization Algorithm

作　　者：王勃[1] 徐静[1] Wang Bo;Xu Jing(Software Teaching and Research Section,Electrical Engineering Department,Shaanxi Instaitute of Technology,Xi'an 710302,China)

机构地区：[1]陕西国防工业职业技术学院,西安710302

出　　处：《计算机测量与控制》2018年第7期195-197,202,共4页Computer Measurement &Control

摘　　要：针对无线网络传感器中如何处理信息服务点中大量的冗余数据、网络运行速度等相关问题;在基于蚁群优化算法的基础上,提出一种BP神经网络的RPROP混合算法;该方法通过在建立系统构架及信息服务点基础上,能够延长BP神经网络的生命周期,加快BP神经网络的收缩速度,能够将网络中信息服务点的重复数据进行有效的合并处理,并及时过滤掉非正常信息服务点的数据,减少数据服务点的能量消耗;仿真结果显示,与普通的蚁群算法相比,该混合算法在训练过程中迭代次数改善明显,解决了BP神经网络的学习、训练时间冗余等问题,同时具有较强的计算、寻优等能力,提高了网络分类正确率和运行的效率;具有一定的实用价值,从而完全能够满足日益增长的无线互联网终端的运行需要。For a large number of redundant data to information service of wireless sensor networks in issues related to network speed.Based on the ant colony optimization algorithm,RPROP proposed a hybrid algorithm of BP neural network.By this method in the establishment of system architecture and information service based on BP neural network,can be extended the life cycle,accelerate the contraction rate of BP neural network,will be able to repeat information service in the network are combined effectively,and timely to filter the non normal information service point data,reduce the data service point of energy consumption.The simulation results show that compared with the conventional ant colony algorithm,the hybrid algorithm in iterative training in the process of solving the number of improved BP neural network learning,training time redundancy,and has strong ability of calculation,optimization,improve the network classification The accuracy and efficiency of operation are of practical value,which can fully meet the needs of the growing wireless internet terminal.

关 键 词：蚁群优化算法 BP神经网络 RPROP混合算法 

分 类 号：TP312[自动化与计算机技术—计算机软件与理论]