资源转移视角下的RCPSP鲁棒资源分配方法

Robust resource allocation method for the RCPSP from a resource transferring perspective

为使项目在工期不确定环境下既能按计划稳定执行又能维持较低的成本,以项目鲁棒性和资源转移成本为优化对象,构建了一个鲁棒资源分配优化模型.引入一种开始时间关键度指标作为项目的解鲁棒性目标,不同于已有研究均采用基于活动的资源流描述,模型定义了基于资源的二元决策变量,以表示某一资源单元在项目活动之间的转移次序.结合遗传算法和模拟退火算法的优点,提出了遗传退火混合智能算法对模型求解,模拟实验结果证明了所提算法在寻优效果和收敛速度方面的优越性.最后通过真实项目案例,进一步验证了模型和算法的实用性与有效性.

In order to achieve schedule stability of a project plan and to maintain comparatively low project costs under activity duration uncertainty,this paper proposes a robust resource allocation optimization model considering both the robustness and resource transfer cost objectives.First,the model adopts a starting time criticality index as a surrogate measure of solution robustness.Second,different from the existing activitybased resource-flow formulations,the model uses resource-oriented binary decision variables to indicate the sequence of activities that will use a particular resource unit.Then,a hybrid intelligent algorithm(GSA)is developed for solving the model,which combines the merits of the genetic algorithm and the simulated annealing algorithm.The results of a simulation experiment demonstrate the superiority of the GSA algorithm in terms of optimization ability and convergence rate.Finally,this article provides a case study of a real project.The comparison results further validate the practicability and effectiveness of the proposed model and algorithm.