基于资源共享的延迟惩罚最小化的多项目调度优化

Scheduling Optimization of Multi-projects Overdue Penalty Minimization Based on Resource Sharing

对可并行进行的多项目调度问题进行定性描述,在共享资源的数量约束下,构建了以多项目延迟惩罚总和为目标函数的优化模型。鉴于问题的强NP-hard属性,为模型设计了迭代循环求解的启发式遗传算法。运用实际案例对研究进行验证,分析关键参数对合同双方项目收益的影响,得到如下结论:与实际进度安排相比,满意进度安排下承包商因延期完工而遭到来自于业主的罚款数额出现了明显的下降;业主的单位延迟罚款费用、共享资源池中的资源数量也会影响承包商对项目的进度安排,并对双方的项目收益产生重要影响。

This paper carries on the qualitative description based on the parallel multi-project scheduling problem.Under the shared-resource constraint conditions,this paper constructs the objective function of the minimization of overdue penalty of multi-projects.Because of the strong NP-hard attribute owned by the model,the authors develop a genetic heuristic algorithm to solve the problem and obtain the satisfied results.A practical example is solved by the heuristic and the influences of the key parameters on the project RCMPSP are analyzed.Based on the results obtained,the following conclusions are drawn：Compared with results under the actual situation, the overdue penalty under the optimal schedule from the client obviously decreased.The overdue penalty cost from the client and the resource quantity from pool influence the project profits remarkably.