考虑VaR的项目伙伴选择问题及其蚁群算法

发布时间：2018-04-11 02:07

本文选题：风险管理 + 伙伴选择　；参考：《东北大学》2014年硕士论文

【摘要】：随着中国经济的快速发展,中国实力的不断增强,政府越来越注重基础设施建设。目前国内水利、交通、通讯、房地产等一大批工程项目蓬勃开展起来,但是工程项目规模大、人员多、耗时长,仅凭一个企业无法独立完成,必须进行工作招标,选择伙伴企业共同完成。由于伙伴企业施工时间的不确定性,使得项目面临拖期风险。因此,对项目进行风险管理是十分必要的。针对项目伙伴选择与拖期风险问题,本文主要做了如下工作：首先综述了项目风险管理与伙伴选择研究的现状及一些相关概念。本文针对伙伴施工时间不确定的项目伙伴选择问题,借助VaR (Value at Risk,在险值)理论的基本概念,建立了以VaR最小为目标的项目伙伴选择不确定性模型；当伙伴施工时间服从正态分布时,根据正态分布的相关概念,将模型转换为确定性模型。其次,针对确定性问题模型,本文设计了4种算法：枚举算法、基本蚁群算法、带信息素上下限的蚁群算法、以及根据问题的特点,在带信息素上下限蚁群算法的基础上做的改进算法——时变参数与关键路径信息引导的蚁群算法。针对三个不同规模的算例,进行了算法参数分析与对比分析,结果表明时变参数与关键路径信息引导的蚁群算法的有效性；并分析了置信水平与变异系数对决策的影响。再次,针对不确定性问题模型,本文设计了全工作蒙特卡洛模拟和关键工作蒙特卡洛模拟两种分析策略,并设计了全工作蒙特卡洛模拟的蚁群算法和关键工作蒙特卡洛模拟的蚁群算法,仿真实验表明全工作蒙特卡洛模拟更为有效。
[Abstract]:With the rapid development of Chinese economy and the increasing strength of China, the government pays more and more attention to infrastructure construction.At present, a large number of domestic water conservancy, transportation, communications, real estate and other engineering projects are booming, but the project is large in scale, has a large number of personnel, and takes a long time. It cannot be completed independently by a single enterprise, so it must be tendered for work.Select partner enterprises to complete the work together.Due to the uncertainty of construction time of partner enterprises, the project faces the risk of delay.Therefore, the project risk management is very necessary.The main work of this paper is as follows: firstly, the current situation of project risk management and partner selection research and some related concepts are summarized.In this paper, aiming at the project partner selection problem with uncertain partner construction time, a project partner selection uncertainty model aiming at the minimum of VaR is established with the help of the basic concept of VaR value at risk theory.When the partner construction time is from normal distribution, the model is transformed into a deterministic model according to the concept of normal distribution.Secondly, for the deterministic problem model, this paper designs four algorithms: enumeration algorithm, basic ant colony algorithm, ant colony algorithm with upper and lower limit of pheromone, and according to the characteristics of the problem,Based on the ant colony algorithm with the upper and lower limits of pheromone, an improved ant colony algorithm with time-varying parameters and critical path information is proposed.The results show that the ant colony algorithm guided by time-varying parameters and critical path information is effective, and the influence of confidence level and coefficient of variation on decision-making is analyzed.Thirdly, aiming at the uncertainty model, this paper designs two kinds of analysis strategies: full working Monte Carlo simulation and critical work Monte Carlo simulation.An ant colony algorithm for full work Monte Carlo simulation and a key work Monte Carlo simulation ant colony algorithm are designed. The simulation results show that the full working Monte Carlo simulation is more effective.
【学位授予单位】：东北大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F283;TP18

【相似文献】