论文部分内容阅读
研究目的:项目成本预测是工程建筑企业进行投标报价的基础,成本预测准确与否最终影响中标结果以及企业利润。工程项目成本的影响因素众多、动态性强,且关系复杂,预测准确性不高。本文在项目成本影响因素识别的基础上,考虑部分因素随时间的变化特征,建立基于时间序列数据的人工神经网络成本预测模型,并结合铁路建筑企业实例对模型进行验证。研究结论:(1)基于成本分析建立的人工神经网络项目成本预测模型能有效预测可投标项目成本,且模型的易用性和客观性更强;(2)考虑时间对部分影响因素的影响,结合时间序列数据修正后的神经网络模型进一步提升了成本预测准确性;(3)本文提出的工程项目投标阶段成本预测模型及输出结果可为工程建筑相关企业的项目投标工作提供决策支持。
Research purposes: Project cost forecasting is the basis for engineering construction enterprises bidding, the accuracy of cost forecasting ultimately affect the successful bidder results and corporate profits. There are many influencing factors of project cost, such as dynamic, complicated relationship and low prediction accuracy. Based on the identification of the influencing factors of project cost, this paper considers the variation characteristics of some factors with time, establishes an ANN-based cost forecasting model based on time-series data, and verifies the model with railway construction enterprises. The conclusions are as follows: (1) The project cost forecasting model based on cost analysis can predict the bidding project cost effectively, and the model is more easy to use and objectivity; (2) Considering the influence of time on some influencing factors, (3) In this paper, the project bidding stage cost forecasting model and the output results can provide decision-making support for the bidding work of project-related enterprises.