【作者】 周婷；

【导师】 刘万琳； 王丽君；

【作者基本信息】 四川农业大学 ， 土木水利硕士（专业学位）， 2023， 硕士

【摘要】 在分布式多项目管理场景中,越来越多的企业面临着因项目数量增加、项目规模扩大以及资源使用紧缺而带来的风险,因此,如何协调多项目资源的分配问题,统筹优化多项目进度安排,是目前有关多项目进度和资源管理急需解决的问题之一。分布式资源约束多项目调度的执行不仅需要各个单项目的独立调度,又需要协调有限的共享资源,因此考虑到多项目环境的复杂性,本文以分布式资源约束多项目调度为研究对象,通过合理分配局部资源(仅在项目内部使用)和全局资源(在项目间共享)来优化多项目进度安排,从而达到多项目收益最大化的目标。本文的主要研究工作如下:首先,在全局资源刚性约束条件下,考虑到刚性资源的有限性,在出现全局资源冲突的时刻点上,采用贪婪选择策略来确定此时刻的活动安排,为其构建以净现值最大化为目标的整数规划模型,并采用遗传算法对该模型进行求解。基于多项目调度问题案例库MPSPLIB中的60个案例进行算法测试,针对关键参数进行了敏感性分析,证明了采用贪婪选择策略来协调活动安排能有效提高项目的净现值。其次,考虑到现实中资源可以通过租赁、购买等方式从市场中获取这一特点,将全局资源刚性约束拓展为柔性约束,在出现全局资源冲突时,从外部获取资源来解决资源不足问题,构建了基于固定资源采购时间尺度下的分布式多项目调度优化模型,并设计遗传-禁忌搜索混合算法求解问题。同样基于MPSPLIB中的案例开展了数值实验,并对关键参数进行了敏感性分析,验证了算法的有效性以及方法的可行性,证明了时间尺度内采购全局资源能有效改善多项目绩效目标。最后,将一个实际案例应用到本文所研究的问题中,分别代入到刚性资源约束分布式多项目调度最大净现值优化模型和柔性资源约束分布式多项目调度最大净现值优化模型中,并采用相应的元启发式算法对问题求解,通过对比分析两种方案下的调度结果。实验结果证明,尽管从外部获取资源会增加项目成本,但是与传统的全局资源刚性约束条件相比,在时间尺度上采购全局柔性资源能显著改善分布式多项目的收益绩效。本文的研究成果有助于企业提前规避项目风险,减少对整体项目进度和经济效益产生的不利影响,从而有效改善分布式多项目绩效目标,能够在实践中为分布式多项目编制合理的进度计划提供决策依据,具有重要的现实意义。更多还原

【Abstract】 In the distributed multi-project management scenario,more and more enterprises are faced with the risks caused by the increase in the number of projects,the expansion of project scale and the shortage of resource.Therefore,how to coordinate the allocation of multiproject resources and coordinate and optimize the schedule arrangement of multi-project is one of the urgent problems concerning multi-project schedule and resource management.The implementation of distributed resource-constrained multi-project scheduling not only requires independent scheduling of each project,but also coordination of limited shared resources.Therefore,Considering the complexity of multi-project environment,this paper takes distributed resource constrained multi-project scheduling problem as the research object,and optimizes the multi-project schedule through reasonable allocation of local resources(only used within the project)and global resources(shared among projects),so as to achieve the goal of maximizing the benefits of multiple projects.The main research work of this paper is as follows:First,under the rigid constraint of global resources,considering the finiteness of rigid resources,the greedy selection strategy is adopted to determine the activity arrangement at the moment when global resource conflicts occur,and the integer programming model aiming at the maximum net present value is constructed for it,meanwhile the genetic algorithm is used to solve the model.Based on 60 cases in the MPSPLIB,the algorithm test and sensitivity analysis of key parameters are carried out.It is proved that using greedy choice strategy to coordinate activity arrangement can effectively improve the net present value of the projectSecondly,considering the fact that resources can be obtained from the market through leasing,purchasing and other means in reality,the rigid constraint of global resources is extended to the flexible constraint.In case of global resource conflicts,resources can be obtained from outside to solve the problem of resource shortage.Building a distributed multi-project scheduling optimization model based on a fixed time scale of resource procurement,and designs a hybrid genetic-tabu search algorithm to solve the problem.Also the numerical experiment was carried out based on the MPSPLIB,and the sensitivity analysis of the key parameters was carried out to verify the effectiveness of the algorithm and the feasibility of the method,then prove that the procurement of global resources within the time scale can effectively improve the performance objectives of multiple projects.Finally,a practical case is applied to the problem studied in this paper,which is substituted into the maximum net present value optimization model of distributed multiproject scheduling with rigid resource constraints and the maximum net present value optimization model of distributed multi-project scheduling with flexible resource constraints,and the corresponding meta-heuristic algorithm is used to solve the problem,then the scheduling results under the two schemes are compared and analyzed.The experimental results show that although obtaining resources from outside will increase the project cost,compared with the traditional rigid constraints of global resources,purchasing global flexible resources on a time scale can significantly improve the revenue performance of distributed multi-projects.The research results of this paper can help enterprises avoid project risks in advance,reduce the adverse impact on the overall project schedule and economic benefits,so as to effectively improve the distributed multi-project performance objectives,which can provide decision-making basis for the establishment of reasonable schedule plans for distributed multi-project in practice,which has important practical significance.更多还原