基于遗传神经网络的军用油库选址优化模型
|
摘要
影响军用油库选址的因素众多且具有不确定性,传统的凭人主观选址和线性规划等方法选址不能充分体现各影响因素的主次成分及相互关系,为了解决该问题,将遗传算法和神经网络相结合,利用德尔斐法建立选址指标体系并进行指标的量化及归一化,将各底层指标的归一化值作为神经网络的输入,将代表选址等级的布尔变量作为神经网络的输出,利用遗传算法来优化神经网络的连接权值和阈值,然后用足够的样本借助Matlab工具训练此模型,通过模型的自适应学习,直到得到能正确表示网络内部特征的那组阈值。实际应用表明,所建模型的操作性和实用性强,为军用油库实际选址提供直接的决策依据。
Because the factors that affect the location of military oil depot are numerous and uncertain,the traditional method of subjective location and linear programming can not fully reflect the primary and secondary components and their relationship. So the combination of genetic algorithm and neural network is suitable for dealing with the above problems. By the method of establishing index system of location,we quantified and normalized index. The normalized values of the underlying indicators were used as inputs to the neural network. The Boolean variable representing the location level is the output of the neural network. The connection weights and threshold used genetic algorithm to optimize the neural network,and then used the Matlab tool to train the model,through the model of adaptivet learning,until it can indicate the internal threshold characteristics of the network group correctly. The practical application shows that the operation and practicability of the model is strong,which provides a direct basis for the actual location of the military oil depot.
Because the factors that affect the location of military oil depot are numerous and uncertain,the traditional method of subjective location and linear programming can not fully reflect the primary and secondary components and their relationship. So the combination of genetic algorithm and neural network is suitable for dealing with the above problems. By the method of establishing index system of location,we quantified and normalized index. The normalized values of the underlying indicators were used as inputs to the neural network. The Boolean variable representing the location level is the output of the neural network. The connection weights and threshold used genetic algorithm to optimize the neural network,and then used the Matlab tool to train the model,through the model of adaptivet learning,until it can indicate the internal threshold characteristics of the network group correctly. The practical application shows that the operation and practicability of the model is strong,which provides a direct basis for the actual location of the military oil depot.
引文
[1]刘鹏飞,査国云.基于GA-PSO的野战给水站选址优化模型[J].四川兵工学报,2014(10):64-67.
[2]董巨为,慕晓冬.基于Arc SDE和Oracle 11g的某装备车选址数据库设计[J].四川兵工学报,2012(8):126-128.
[3]朱来辉,曹炜.基于FAHP的民用机场选址方案优选[J].四川兵工学报,2012(6):117-119.
[4]李海峰,康中敏.基于混合遗传神经网络的电动汽车充电站最优选址的研究[J].四川电力技术,2012(4):49-52.
[5]张经博,郭凌.基于遗传算法优化BP神经网络的供暖系统热负荷预测模型[J].四川兵工学报,2014(4):152-156.
[6]赵建春,叶丽娜.遗传算法在故障诊断中的应用[J].四川兵工学报,2013(3):132-134.
[7]李晔.基于排序选择的改进遗传算法优化的神经网络控制[J].四川兵工学报,2010(9):117-119.
[8]邱国庆,何雷.基于神经网络的野战给水能力评价模型[J].四川兵工学报,2009(2):32-35.
[9]黄国栋.基于遗传神经网络的短期负荷预测研究[J].中国电力教育,2014(2):261-262.
[10]胡锦宏.基于遗传神经网络的煤层气含量预测[J].绿色科技,2015(2):260-263.
[11]杨晓段,李元左.非战争军事行动装备保障资源运输方式选择模型[J].四川兵工学报,2013(12):35-40.
[2]董巨为,慕晓冬.基于Arc SDE和Oracle 11g的某装备车选址数据库设计[J].四川兵工学报,2012(8):126-128.
[3]朱来辉,曹炜.基于FAHP的民用机场选址方案优选[J].四川兵工学报,2012(6):117-119.
[4]李海峰,康中敏.基于混合遗传神经网络的电动汽车充电站最优选址的研究[J].四川电力技术,2012(4):49-52.
[5]张经博,郭凌.基于遗传算法优化BP神经网络的供暖系统热负荷预测模型[J].四川兵工学报,2014(4):152-156.
[6]赵建春,叶丽娜.遗传算法在故障诊断中的应用[J].四川兵工学报,2013(3):132-134.
[7]李晔.基于排序选择的改进遗传算法优化的神经网络控制[J].四川兵工学报,2010(9):117-119.
[8]邱国庆,何雷.基于神经网络的野战给水能力评价模型[J].四川兵工学报,2009(2):32-35.
[9]黄国栋.基于遗传神经网络的短期负荷预测研究[J].中国电力教育,2014(2):261-262.
[10]胡锦宏.基于遗传神经网络的煤层气含量预测[J].绿色科技,2015(2):260-263.
[11]杨晓段,李元左.非战争军事行动装备保障资源运输方式选择模型[J].四川兵工学报,2013(12):35-40.