基于贝叶斯网络的施工项目成本动态预测方法研究

发布时间：2017-12-31 01:05

本文关键词：基于贝叶斯网络的施工项目成本动态预测方法研究　出处：《昆明理工大学》2014年硕士论文　论文类型：学位论文

【摘要】：随着国民经济快速发展,建筑施工行业进入了跨越式发展的新时代,国内外大量建筑施工企业加入到市场化竞争。建筑施工行业面对复杂多变的市场环境,竞争激烈、利润低薄,生存压力巨大,因此大多数建筑施工企业将管理的重点放在成本管理工作上。而建筑施工企业成本管理首当其冲的工作是成本预测,它是施工项目编制成本计划、实施成本管控的首要环节,同时也是施工企业投标阶段进行项目决策的主要依据,对于提升和改善施工企业的生存发展能力以及市场竞争能力,具有举足轻重的影响力。目前国内外成本预测的方法主要存在以下问题：(1)部分施工企业项目成本意识淡薄,未形成系统的成本管理体系。(2)成本预测方法落后,成本管理水平偏低。(3)片面对待成本管理,在成本预测过程中忽视了质量、工期、安全及标准化建设与成本之间的对立统一关系。国内外学者采用神经网络方法、灰色理论、挣值分析法、模糊理论、遗传算法、影响因素分析、工程量清单计价分析、多元线性回归法、地质预测方法、贝叶斯网络和传统风险评估法等多种模型和方法进行了施工项目成本的预测及研究,开展施工项目成本预测的多方面探索,获得了很大成效,但这些方法仍然存在以下几个方面的缺陷：(1)以往的研究均只能实现对施工项目成本的静态预测,而不能实现任意时点施工成本的动态预测；因此,这些成本预测方法的应用被局限在施工投标阶段,而不能全面应用于指导施工项目成本风险的动态控制。(2)研究的方法大部分没有考虑成本的形成过程多个影响因素之间复杂的交互作用关系,以及这种复杂关系影响最终施工项目成本的机理；因此很难用于指导和改善成本控制过程。本文将ISM方法与贝叶斯网络相结合,分析研究施工项目成本影响因素,确定各影响因素对造价构成分项的影响程度和相互关系,建立ISM解释结构模型,基于ISM模型进行贝叶斯网络分析预测,构建了施工项目的成本动态预测模型,并通过工程案例对该模型进行了验证。本文的研究主要有以下优势和成果： (1)基于贝叶斯网络的施工项目成本预测方法思路清晰,方法简便,在实际工作中也很容易操作,反映出来的结果与实际情况吻合度良好,可靠性较高,而且能够即时动态预测项目实施过程中的特定成本,因此,在施工项目成本管理领域具有较好的应用前景。 (2)通过基于贝叶斯网络的施工项目成本预测模型的正向分析可以迅速求解项目的预期成本,而通过贝叶斯网络模型的反向分析可以迅速求出项目成本风险的致险因子,以此为依据可以实现施工项目成本的动态预警,并未项目的成本管理与成本控制提供有效的指导。
[Abstract]:With the rapid development of the national economy, the construction industry has entered a new era of leapfrog development, a large number of domestic and foreign construction enterprises to join the market competition. The construction industry faced with complex market environment, fierce competition, low profits thin, great pressure to survive, so most of the construction enterprise management will focus on cost management work. The construction enterprise cost management work is the cost forecast, it is the preparation of construction project cost plan, the primary aspects of the implementation of cost control, but also the main basis for construction project decision-making enterprises bidding stage, to enhance and improve the construction enterprise's survival and development ability and market competition ability, has a great influence methods at home and abroad. The cost of the forecast mainly has the following problems: (1) part of the construction enterprise project cost consciousness (2) the cost prediction method is lagging behind and the cost management level is low. (3) one-sided cost management is neglected in the cost prediction process, which ignores the unity of opposites between quality and duration, safety and standardization construction and cost.
Scholars at home and abroad by using neural network method, grey theory, earned value analysis method, fuzzy theory, genetic algorithm, factor analysis, BOQ valuation analysis, multiple linear regression method, geological prediction method, Bayesian network and traditional risk assessment method of various models and methods of prediction and research of the construction project cost, explore in many ways to carry out the prediction of construction project cost, achieved great results, but these methods still exist defects in the following aspects: (1) previous studies were only used to predict the static state of the construction project cost, and can not achieve dynamic prediction at any time point for the construction cost; therefore, the application of prediction method of these costs is limited in the construction bidding stage, dynamic control and not fully applied to guide the construction project cost risk. (2) research methods most did not test the formation process of considering cost It is difficult to guide and improve the cost control process due to the complex interaction between multiple factors and the mechanism of the complex relationship affecting the final construction project cost.
In this paper, the ISM method and Bias network combined with the analysis of the influencing factors of the construction project cost, to determine the influence factors and the relationship between the degree of breakdown of the cost, the establishment of ISM structural model, Bias network analysis and prediction based on ISM model, the construction cost of construction project dynamic prediction model, and the model was verified by the case study. This research mainly has the following advantages and achievements:
(1) Bayesian network construction project cost forecast method based on clear thinking, the method is simple, in practice is also very easy to operate, reflected the results agree with the actual condition of good, high reliability, but also the specific cost, real-time dynamic prediction in the process of project implementation, therefore, has a good application prospect in the field of cost the management of construction projects.
(2) can be expected to quickly solve the project cost through positive analysis prediction model based on Bayesian network construction project cost, and the reverse of the Bayesian network model analysis quicklysolute project cost risk risk factors, on this basis can realize dynamic early warning construction project cost, cost management and cost of the project is not to provide effective guidance.

【学位授予单位】：昆明理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TU723.3

【参考文献】