设备状态监测平台数据驱动的制造商备件生产与库存策略研究

随着设备状态监测平台的大规模部署和应用，基于状态监测数据的备件支持服务已成为我国装备制造企业的重要利润来源。相比用户各自为战的传统备件库存管理，制造商主导的备件支持服务具有更准确的需求预测、生产与库存联合控制、数据驱动的高频次决策等诸多优势。作为备件支持服务的两大核心驱动，状态监测数据实时反映个体部件的劣化过程，刻画了备件需求产生的微观基础；设备部件结构是备件支持服务的技术输入，决定了备件需求的种类和数量。制造商只有深入认知这两方面特征，才能实现精准的备件需求预测和科学的生产库存决策。根据状态监测数据的特征（监测指标、连续性和推断作用），结合三类设备部件结构（单部件系统、同种多部件系统和多种多部件系统），本项目拟系统梳理备件支持服务的常见应用场景，全面研究数据驱动的制造商备件生产与库存最优控制策略。本项目能够丰富数据驱动的库存理论研究，同时对指导相关企业生产与库存管理实践有着重要现实价值。

With the popularity of equipment monitoring platforms, spare parts support service based on equipment state data has become an important business in equipment manufacturing industry. Compared with spare parts inventory managed by users, spare parts support service has many advantages, such as better demand forecast, joint production and inventory control, data-driven decision making with higher frequency, etc. Equipment state data and equipment structure information are two critical inputs in offering high-standard spare parts support service. The former updates the current status of components' deterioration and the latter determines material requirements of maintenance. We classify different application scenarios by features of equipment state data, such as monitoring measures, data continuity and information inference. We study three types of equipment structure, including equipment with a single component, equipment with multiple components of one kind, and equipment with multiple components of multiple kinds. For each scenario, we will investigate its optimal policy of joint production and inventory control driven by equipment state data. Our research would enrich data-driven inventory control theory and provide analytical tools for practitioners in equipment manufacturing industry.