摘要
针对多属性决策中的高维、非线性问题,提出一种基于粗糙集和粒子群优化神经网络的智能多属性决策方法.该方法利用粗糙集对多属性决策问题的条件属性进行约简,利用粒子群算法训练神经网络的权重和阈值形成粒子群优化神经网络模型,约简后的属性数据进入粒子群优化神经网络的智能决策系统.实证结果表明,该方法具有较好的泛化能力,与标准支持向量机、遗传神经网络等方法相比,该方法具有一定的优势.
For solving the high dimensional and nonlinear problems of multiple attribute decision making （ MADM ）, in this paper, an intelligent method based on rough sets（ RS ） ,particle swarm optimization（ PSO ） and artificial neural network （ ANN ）, RSPSOANN ,is proposed. In this hybrid approach,RS is used for attribute selection in order to reduce the model complexity of ANN and improve the speed of ANN, PSO is used to train the weights of ANN to constitute a PSOANN model, and then the reduced data is introduced into PSOANN to obtain the results of decision making. The empirical results reveal that RSPSOANN method has understanding forecasting ability. Compared with the standard SVM and GAANN （ training artificial neural network with genetic algorithm）, RSPSOANN has some superiority in predicting accuracy.
出处
《小型微型计算机系统》
CSCD
北大核心
2009年第5期920-923,共4页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(70471074)资助
关键词
多属性决策
粗糙集
粒子群算法
人工神经网络
multiple attribute decision making
rough sets
particle swarm optimization
artificial neural network
