不确定环境下资源受限多项目调度与控制

发布时间：2018-02-28 01:07

本文关键词： 多项目调度不确定资源受限遗传算法系统动力学　出处：《西南交通大学》2017年硕士论文　论文类型：学位论文

【摘要】：社会经济的不断发展催生了大量的项目建设需求,多项目管理已成为企业常态。然而由于多项目情况下管理界面较大,且存在子项目之间的有限资源争夺关系,导致多项目调度成为了研究和实践中的重难点问题。另外,由于国内城市化进程加速导致项目环境越来越复杂,为项目实践增添了很多不确定性因素,这些不确定性因素的存在进一步增大了项目调度难度,使得工程延期、成本超支等情况时有发生。不确定环境下多项目调度问题是调度领域亟待解决的问题。针对这一问题,本文采用文献分析法、定性分析法、定量分析法等方法对不确定环境下多项目调度与控制问题进行研究,提出了 "计划-控制"两阶段模型。首先,构造了基于遗传算法的资源受限多项目调度模型。该算法采用精英+排序选择方式保证了最优个体不被排除的同时,也避免了算法过早收敛;优先关系保留交叉算子保证了子代依然满足紧前关系约束,解决了算法产生不可行解的问题。紧接着,构建了基于系统动力学的不确定环境下多项目控制模型。该模型共包含项目进度、项目表现、项目资源、工期变化等四个子系统,分别模拟了项目实践中的重要反馈结构,其中进程子系统中包含了各任务的紧前关系约束和资源约束,实现了项目调度计划和项目控制的有效对接。通过算例仿真与分析,表明了该"计划-控制"两阶段模型进行不确定环境下资源受限多项目调度与控制的有效性。首先构造了一个包含3个子项目、36个实任务的多项目算例;然后运用遗传算法分别求解算例在无资源约束、含资源约束、含项目权重系数等情况下的调度计划;最后运用系统动力学模型进行不确定环境下多项目控制仿真。仿真结果表明,"计划-控制"两阶段模型能够有效应对不确定环境下资源受限多项目调度问题。
[Abstract]:The continuous development of social economy has given birth to a large number of project construction needs, and multi-project management has become the norm of enterprises. However, due to the large management interface in multi-project situations, there is a limited resource contention relationship between sub-projects. As a result, multi-project scheduling has become a very important and difficult problem in research and practice. In addition, due to the increasing complexity of the project environment due to the acceleration of the domestic urbanization process, a lot of uncertainties have been added to the project practice. The existence of these uncertain factors further increases the difficulty of project scheduling and makes the project delay and cost overrun happen from time to time. The multi-project scheduling problem in uncertain environment is a problem to be solved urgently in the field of scheduling. In this paper, the methods of document analysis, qualitative analysis and quantitative analysis are used to study the problem of multi-project scheduling and control in uncertain environment, and a two-stage model of "plan-control" is proposed. A resource-constrained multi-project scheduling model based on genetic algorithm (GA) is constructed, which adopts elite sorting and selection method to ensure that the optimal individual is not excluded and to avoid premature convergence of the algorithm. The precedence relation preserves the crossover operator to ensure that the offspring still satisfies the constraint of the compact front relation, and solves the problem that the algorithm produces the infeasible solution. A multi-project control model in uncertain environment based on system dynamics is constructed. The model consists of four subsystems: project schedule, project performance, project resource and duration change, respectively, and simulates the important feedback structure in project practice. The process subsystem includes the constraints of each task and the resource constraints, and realizes the effective docking of the project scheduling plan and the project control. The effectiveness of the "plan-control" two-stage model for resource constrained multi-project scheduling and control in uncertain environments is demonstrated. First, a multi-project example consisting of 3 subprojects and 36 real tasks is constructed. Then genetic algorithm is used to solve the scheduling plan in the case of no resource constraint, including resource constraint and item weight coefficient respectively. Finally, the system dynamics model is used to simulate the multi-project control in uncertain environment. The simulation results show that the "plan-control" two-stage model can effectively deal with the resource-constrained multi-project scheduling problem in uncertain environment.
【学位授予单位】：西南交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F224;F272

【相似文献】