公路交通项目虚拟集成投资估算决策技术研究

发布时间：2019-01-19 17:09

【摘要】：工程项目前期的投资估算是优选方案以及资金筹措的依据,同时对工程项目总成本的控制有着至关重要的作用。以往投资估算方法的简单滞后性导致估算误差较大、结果不够准确,因而寻找贴合工程实际、更为科学有效的投资估算方法来保证工程项目投资估算的准确度是急需解决的问题。本文以公路交通项目为例,以全生命周期显著性造价理论(WLCS)和已完类似工程为基础,通过对公路工程建设项目的特点分析,依据拥有训练样本的不同情况建立相应的非线性投资估算模型来拟合工程项目造价与其各影响因素之间的非线性关系,从而对公路交通项目进行造价预测。首先利用粗糙集(RS)的属性约简特性来挖掘工程数据,提取建设项目的工程特征,克服以往寻找有效工程特征方法的主观性,从而证明该方法的科学有效性。面对拟建工程项目的训练样本数量的不同,分别采用不同的非线性投资估算方法。若训练样本的数量一定,采用模糊聚类FC估算方法对拟建项目进行造价估算,通过实例验证该方法是有效可行的。而面对大量训练样本的情况,则采用智能集成估算方法,主要包括粗糙集—神经网络(RS-BP)估算法、蚁群—神经网络(ACO-BP)估算法以及粒子群—径向基网络(PSO-RBF)估算法。RS-BP估算法是用粗糙集首先对网络输入变量作约简预处理,然后利用网络进行训练预测。ACO-BP与PSO-RBF估算法是利用群智能对神经网络进行优化,从而得到智能集成估算方法。通过案例仿真证明该算法更加贴合工程实际,大大加快训练速度,降低误差,提高工程造价预测精度,体现了该算法的科学性与优越性。在以上方法基础上,运用虚拟技术建立投资方案虚拟可视化模型,使投资方案直观形象展现在决策者面前。
[Abstract]:The investment estimation in the early stage of the project is the basis for the optimal selection of the project and the raising of funds. At the same time, it plays an important role in the control of the total cost of the project. In the past, the simple lag of investment estimation method resulted in large estimation error and inaccuracy of the result, so it is practical to find a fitting project. More scientific and effective investment estimation method to ensure the accuracy of project investment estimation is an urgent problem. This paper takes the highway transportation project as an example, based on the life-cycle significant cost theory (WLCS) and similar projects, through the analysis of the characteristics of highway construction projects. According to the different conditions with training samples, the corresponding nonlinear investment estimation model is established to fit the nonlinear relationship between the project cost and its influencing factors, so as to predict the cost of highway traffic projects. Firstly, the attribute reduction characteristic of rough set (RS) is used to mine engineering data, to extract the engineering features of construction projects, to overcome the subjectivity of the previous methods of finding effective engineering features, and to prove the scientific validity of this method. In the face of the different training samples of the proposed project, different nonlinear investment estimation methods are used. If the number of training samples is constant, the method of fuzzy clustering FC estimation is used to estimate the cost of the proposed project. The example shows that this method is effective and feasible. In the case of a large number of training samples, intelligent ensemble estimation method is used, including rough set neural network (RS-BP) estimation method. Ant Colony Neural Network (ACO-BP) estimation and Particle Swarm Radial basis Network (PSO-RBF) estimation. The RS-BP estimation method uses rough set to preprocess the input variables of the network. ACO-BP and PSO-RBF estimate method is to optimize the neural network by using swarm intelligence, and then the intelligent integrated estimation method is obtained. The simulation results show that the algorithm is more suitable to the engineering practice, greatly speeds up the training speed, reduces the error, and improves the accuracy of the project cost prediction, which reflects the scientific nature and superiority of the algorithm. On the basis of the above methods, the virtual visualization model of investment scheme is established by using virtual technology, so that the visual image of investment scheme can be displayed in front of decision makers.
【学位授予单位】：石家庄铁道大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F542

【参考文献】