摘要
进度计划的稳定性对不确定环境下项目的顺利实施具有重要影响。从活动自由时差效用函数的新视角度量调度方案的鲁棒性,进而,同时考虑项目工期和鲁棒性两个目标函数,构建了基于时差效用的双目标资源约束型鲁棒性项目调度优化模型(BORRCPSP)。针对模型的NP-hard和多目标组合优化特征,设计一种调整的快速非支配性排序多目标遗传进化算法(NSGA-II)求解模型。基于PSPLIB中的480个标准算例进行大规模数值仿真测试。采用获得Pareto最优解集合的算例比率、与最优工期的偏差以及Pareto最优解集合的规模和计算时间作为4个评价指标,实验数据的统计结果验证了BORRCPSP模型和调整的NSGA-II算法的有效性。结论表明,BORRCPSP的Pareto最优解集合可以给项目经理综合考虑项目工期和进度计划的稳定性提供定量化决策依据。
The stability of a schedule has a great uncertain conditions. First, this paper measures influence on the successful execution of the project under the robustness of a schedule from a novel perspective of utility functions of the free floats of an activity, and the model of a bi-objective robust resource-constrained project scheduling problem with an utility function of activity floats （BORRCPSP） is constructed, which simultaneously considers two performances （project makespan and robust measurement）. Second, aimed at the NP-hard and multi-objective combinational characteristics for the BORRCPSP, an adjusted fast elitist non-dominated sorting genetic algorithm （NSGA-II） is developed to solve the model. Next, an extensive numerical experiment is designed and conducted based on the set of 480 benchmark instances from the PSPLIB. Four indexes are adopted in the experiment, including the ratio of obtained set of optimal Pareto solutions for all instances, the deviation from the optimal makespan, the size of set of optimal Pareto solutions, and the CPU time. The statistical values of the experimental results verify the effectiveness of the BORRCPSP and the adjusted NSGA-II algorithm. The conclusion shows that the set of optimal Pareto solutions from the BORRCPSP can provide project managers with quantitative decision basis in order that they can synthetically think about the makespan and the robustness of a schedule for a project,
作者
张静文
周杉
乔传卓
ZHANG Jingwen, ZHOU Shan, QIAO Chuanzhuo(School of Management, Northwestern Polytechnical University, Xi＇an 710072, Chin)
出处
《系统管理学报》
CSSCI
CSCD
北大核心
2018年第2期299-308,共10页
Journal of Systems & Management
基金
国家自然科学基金资助项目(71572148)
中国博士后科学基金资助项目(2015M580875
2016T90947)
航空科学基金资助项目(2015ZG53080)
陕西省博士后基金资助项目(2017BSHYDZZ22)
社会科学基金资助项目(2014P23)
西北工业大学研究生创新基金资助项目(Z2017055
Z2018037)
关键词
鲁棒性
双目标
调度方案
效用函数
约束型
调度优化
utility functions of activity floats
robust measure
bi-objeetive
resource-constrained project scheduling problem （RCPSP）
non-dominated sorting genetic algorithm （NSGA-II）
作者简介
张静文（1976-），女，博士，教授，博士生导师。研究方向为项目调度优化管理。E-mail：zhangjingwen@nwpu．edu．cn
