基于数字型设计结构矩阵-遗传算法的项目管理方法研究

Research on Project Management Method Based on RDSM-GA

项目管理长期存在因项目任务数量繁多、任务关系复杂导致并行效率低下的问题,传统的项目管理方法不能直观表示任务关系,无法对风险耦合较高的任务进行识别、解耦,产生项目进度与成本风险损失,严重影响任务并行实施效果。针对此问题,基于项目目标拆解实现目标所需任务,提出利用数值设计结构矩阵表达项目任务关系,构建风险概率矩阵与风险损失矩阵对项目总体预计风险损失值进行评估。以某空间科学任务并行设计项目流程优化为实证案例,使用遗传算法求解项目预计风险损失值最低流程,提出科学目标满足研究任务组、有效载荷相关任务组、地面站系统任务组和发射场相关任务组等四大项任务组宏观并行、微观串行和局部概率任务反馈的流程,结果显示反馈任务最少优化后风险时间损失和经济损失分别减少7.4%、8.8%,风险损失最低优化后风险时间损失和经济损失分别减少3.0%、5.0%,与实际项目开展结果相符,证明基于数字型设计结构矩阵-遗传算法(RDSM-GA)寻优方法的可行性。

Project Management exists a long-standing problem of low parallel efficiency due to the large number of project tasks and complex task relationships,the traditional methods cannot visually express task relationships,and cannot identify or decouple project tasks with high risk coupling,due to extra cost and seriously affecting concurrent engineering.In order to solve this problem,dismantling tasks based on project objectives,this paper uses the numerical design structure matrix to express the task relationship of the project,and constructs the risk probability matrix and risk loss matrix to evaluate the overall expected risk loss value of the project.Taking the optimization of a parallel design project process for a spatial science task as an empirical cases,the paper uses the method of genetic algorithm to calculate the process of the minimum expected risk loss value of the project,proposes a flow of macroscopic parallel,microserial,and local probabilistic task feedback for the four task groups,including scientific objectives meet the research task group,payload-related task group,ground station system task group,and launch field-related task group,the result shows that the risk time loss and the risk economic loss are reduced by 7.4%and 8.8%respectively after choosing the minimum number of feedback tasks as the optimization object,and the risk time loss and the risk economic loss are reduced by 3.0%and 5.0%respectively after using the method constructed in seeking minimum of risk lost,which are consistent with the results of the actual project.The results demonstrate the feasibility of risk design structure matrix based genetic algorithm(RDSM-GA).