摘要
人工神经网络是一门崭新的信息处理学科,在工程领域已有广泛应用。文章分析了人工神经网络BP模型在工程投标报价风险金预测中存在的问题,认为BP训练存在平坦区域,容易陷入局部极小而得不到全局最优解,在工程投标报价风险测算中,GA算法+BP模型与BP算法相比,无论在收敛速度上,还是在精度上都取得了很好的效果,预测值的精度足以满足工程要求。
Artificial Neural Network (ANN), as a new kind of intelligent information processing theory, has been widely applied in engineering. BP model of ANN has some shortcomings in forecasting project bid- ding pricing risk cost, i.e. there is a flat zone in BP model training and it is apt to get a local minimum other than a global optimal solution. GA arithmetic plus BP model is better than BP model concerning convergence speed and calculation precision. The forecast precision is sufficient for engineering demand.
出处
《石油工程建设》
2009年第1期66-68,共3页
Petroleum Engineering Construction
关键词
人工神经网络
BP网络
GA算法
预测模型
风险金
Artificial Neural Network
BP Network
GA arithmetic
forecast model
risk cost
