Abstract: There are different human resources with different skills in the service scheduling of Complex Product Maintenance,Repair and Overhaul (MRO).To solve the problem that the actual execution period of the task will change with the different assigned schemes,the matching relation of "task-skill-personnel" was established by combining multi-skill labor scheduling with personnel learning effect.A stochastic chance constrained programming mathematical model was established under uncertain resource scheduling time parameters with objectives of minimizing service time,minimizing personnel redundancy and maximizing the cost performance index of the resources.A hybrid intelligent algorithm based on stochastic simulation,BP neural network and NSGA-II was proposed to solve the problem.The effectiveness and applicability of the proposed method were verified by comparative experimental cases with different parameter characteristics and sensitivity analysis.

Key words: maintenance,repair and overhaul service resource scheduling, multi-skilled human resources, temporal uncertainty, stochastic chance-constraint, hybrid intelligent algorithm

摘要： 针对复杂产品维护、维修、大修(MRO)服务资源调度中存在人力资源多技能异质而导致任务实际执行工期随着指派方案的不同而变化的问题,将多技能工调度与人员学习效应相结合,建立“任务技能人员”的匹配关系。以最小化服务完成时间、最小化人员冗余和最大化资源性能成本比为目标,构建资源调度时间参数不确定条件下的随机机会约束规划数学模型。并提出一种基于随机模拟、BP神经网络和改进NSGA-II的混合智能算法进行求解。通过具有不同参数特征的对比实验案例和敏感性分析,验证了所提模型和算法的有效性和适用性。

关键词: MRO服务资源调度, 多技能人力资源, 时间不确定性, 随机机会约束, 混合智能算法