建立了一套系统地选择勘察承包商的架构和体系。首先,基于工程实际,优化提出了大型工程建设项目勘察承包商的输入选择属性,分为技术属性集和商务属性集。其次,将人工神经网络技术应用于勘察承包商的实际勘察产出预测,建立了科学合理的神经网络结构,对勘察承包商实际勘察产出进行预测。最后在实际产出预测基础上,确定如何选择勘察承包商。在提出算法的基本思想和步骤后,利用Matlab作为实验工具,选用实例进行了预测和选择。实验结果显示,模型具有自学习能力,有较高的预测正确率,能够用来进行选择勘察承包商。
The system of selecting the investigation contractor is established in the paper. At first, according to engineering project, the input selection properties of investigation contractor is proposed and optimized, which is categorized into technical set and business set. Moreover, artificial neural network is applied to the investigation' s output prediction, and an appropriate neural network structure is established to predict the future output of the contractor. Finally, the investigation contractor is selected based on prediction of accurate output of the contractor. After putting forward the algorithm' s basic idea and steps, it uses Matlab as experimental tool and selects some cases of model to predict and select. The experiment shows that the model has self-learning ability and high accuracy of prediction, and it can be used to select contractors.