基于改进BP神经网络的工程建设项目评标方法研究

针对评标工作中专家的主观性、随意性和倾向性所带来的偏差,运用改进BP算法,构建了建设工程项目评标的神经网络模型。该模型能够较好地适应各影响因素之间的非线性关系,并通过附加动量法的自学习、自适应能力,提高了收敛速率,从而避免了局部极小值,提高了精度。通过仿真试验,验证了该模型的科学性和有效性。

Research on Bid Evaluation Method for Construction Projects Based on Improved BP Neural Network

According to the deviation caused by experts' subjectivity and arbitrary bias, builds a neural network model for construction project bidding with improved BP algorithm. This model can adapt to the nonlinear of each influencing factor and use additional momentum algorithm with self-learning and self-adapt to improve convergence rate and precision. So it can avoid local minimum of the bidding result. Through the simulated test, verifies that the model is scientific and effectiveness.