服务型制造系统中维修与生产计划的协同调度机制研究

The manufacturing industry is experiencing the transform from "production-oriented manufacturing" to "service-oriented manufacturing". New phenomena and factors appear and challenge the traditional manufacturing approach and theory. For example, customers are allowed and encouraged to participate in the process of manufacturing in the service-oriented manufacturing. Another important one is the outsourcing of producer service, such as the preventive maintenance. All of these demand new operation management and scheduling theory and approach urgently. The project studies the jointly scheduling problem of production and maintenance policies in the service-oriented manufacturing system. To begin with, we discuss the feature of the service-oriented manufacturing. A customer behavior model of participating in manufacturing is built and the correlation between manufacturing and customer behavior is analyzed, on the basis of which a production scheduling model integrating customer behavior is built. A sequential decision based approach is presented. Secondly, the outsourcing decision model of maintenance service is formulated as a linear programming model. A preventive maintenance plan optimization model with the consideration of outsourcing is built as a programming model and solved by genetic algorithm. Thirdly, the mechanism between production and maintenance is discussed and a model is formulated to jointly schedule production and preventive maintenance. In the sequential scheduling approach, a method combining genetic algorithm and local search motived by the distribution of electronic layer outside the atomic structure is proposed. An evaluation method integrating analytic hierarchy process with the fuzzy mathematics is put forward to select the best compromise solution from the Pareto solutions. Finally, simulation experiments are carried out with emplant software to verify and analyze the models and the approaches. A scheduling prototype system is designed and developed in order to put the above research conclusion into practice. The project belongs to applied basic research. The research result will lay foundation for the production operation management and control in service-oriented manufacturing systems. Therefore the research on the jointly scheduling problem of production and maintenance policies in the service-oriented manufacturing system is of great importance both in theory and in practice.

现代制造业正从"生产型制造"向"服务型制造"转变，顾客参与和业务外包是服务型制造的两大特征，由此产生的新问题成为目前制造业迫切需要解决的重要问题。本项目围绕服务型制造系统中维修和生产计划的协同调度问题开展研究。系统分析上述特征，建立顾客参与的行为模型，揭示其对制造过程的影响规律，从而建立基于顾客参与的生产调度模型，提出调度优化方法；建立维修外包决策模型并提出优化方法，在此基础上，建立维修计划优化模型并提出优化方法；最后分析维修与生产的关联关系，建立服务型制造系统中维修和生产计划的协同调度优化模型，在序贯调度优化框架内，集成遗传算法和基于能级的局部搜索算法提出优化方法，设计Pareto解评价方法；最后仿真分析，开发原型系统并验证。本项目属于应用基础研究，将形成一套针对服务型制造系统的维修和生产计划的协同调度理论和方法，能为服务型制造系统的生产运作提供理论支撑，具有重要的理论意义和应用价值。

随着云计算、大数据等先进科学技术的出现，现代制造业正从“生产型制造”向“服务型制造”转变，原来的生产管理和控制方法及手段都面临着新的挑战，如何科学高效地对服务型制造系统进行管理和控制成为摆在决策层面前迫切需要解决的重要问题。本项目从分析服务型制造系统的特征出发，总结出“顾客参与”和“业务外包”是服务型制造系统的两大特征，然后以柔性作业车间生产环境作为应用场景展开研究。（1）建立了柔性作业车间调度的优化模型，提出了求解该问题的启发式规则库；提出了超启发式遗传算法、改进的细菌觅食算法、双机制头脑风暴算法、超启发式文化基因算法、精英量子进化算法和变邻域改进遗传算法求解该问题。（2）对比分析了现有预维护模型，选择了更具有通用性的扩展的混合PM模型作为预维修策略模型；采用威布尔分布模型建立了机器故障模型；结合预维护模型和机器可靠性函数，得到了预维护任务的可靠性模型。（3）建立了维修与生产计划的协同调度模型；提出了“先维护后调度”、“先调度后维护”、“生产与维护同步调度”的柔性作业车间维修与生产计划协同调度策略；对比分析了三种策略，数值实验结果证明“生产与维护同步调度”策略表现最优秀；在此基础上，提出了一种基于役龄模型的同步协同优化策略，进一步优化了协同调度策略并进行了实验验证。（4）建立了柔性作业车间动态调度优化模型；提出了右移重调度策略、完全重调度策略、部分重调度策略和基于关联树和机器路径的部分重调度策略，数值实验结果表明基于关联树和机器路径的部分重调度策略是求解该类问题的最优策略。（5）基于B/S架构开发了调度原型系统，将上述理论研究成果融入到原型系统里，根据生产计划自动产生生产调度结果和维修计划的调度结果。本项目属于应用基础研究，形成了一套针对服务型制造系统的维修和生产计划的协同调度理论和方法，能为服务型制造系统的生产运作提供理论支撑，具有重要的理论意义和应用价值。