投资估算是项目建议书和可行性研究报告的重要组成部分,是项目投资决策的主要依据之一.项目估算准确性直接影响设计概算与施工图预算的编制.估算的影响因素与其结果之间存在复杂的非线性映射关系,传统数学方法用于解决非线性映射问题时具有很大的局限性.为了提高投资估算的准确性,基于遗传 BP 神经网络,对误差反向传播机理进行深度分析,引入灰色系统理论,得到样本数据之间变化规律以及弱化适应度函数值的波动性,建立了灰色遗传神经网络预测模型.并对已有样本数据进行仿真,结果显示灰色遗传神经网络模型误差均值为 1.54%,优于 GM(1,1)预测模型、标准 BPNN模型、GA-BPNN模型.验证了文中所建立的模型在工程估价中的有效性,对工程建设成本控制具有一定的实际意义.
COST ESTIMATION AND PREDICTION OF CONSTRUCTION PROJECTS BASED ON GRAY GENETIC NEURAL NETWORK
Investment estimation is an important part of project proposals and feasibility study reports,and is one of the main bases for project investment decisions.The accuracy of project estimates directly affects the preparation of design estimates and construction drawing budgets.There is a complex nonlinear mapping relationship between the estimated influencing factors and their results,and traditional mathematical methods have great limitations when used to solve nonlinear mapping problems.In order to improve the accuracy of investment estimation,based on the genetic BP neural network,the error backpropagation mechanism is analyzed in depth,the gray system theory is introduced,the variation law between sample data and the fluctuation of the weakened fitness function value are ob-tained,and the gray genetic neural network prediction model is established.The simulation results show that the average error of the gray genetic neural network model is 1.54%,which is better than the GM(1,1)estimation model,standard BPNN model and GA-BPNN model.This paper verifies the effectiveness of the model established in this paper in project valuation,which has certain practi-cal significance for engineering construction cost control.
project investment estimatenonlinear mappinggrey systems theorygenetic neural networks
NETL
NSTL
万方数据
