复杂生产环境下的随机客户订单调度问题研究

Customer orders are naturally associated with multiple product types, each with a specific quantity corresponding to order requirements. Customer order scheduling refers to the practice of how to efficiently schedule the limited resources so as to meet the demand of customer orders. With the ever-increasing competition worldwide, it is imperative for manufacturers to develop more advanced customer order scheduling strategies to ensure competitive advantages. In order to achieve this goal, three main problems have to be tackled: 1) the uncertainty of customer orders about arriving time, product types and workload, 2) the significant variance among customer orders, 3) the coordination of every element in a complicated manufacturing environment. This research intends to establish a more realistic model of customer order scheduling problems when orders arrive dynamically and schedules have to be performed online. The optimality properties of the new model will be explored. Based on these properties, a set of effective algorithms will be developed which can be applied to a variety of order scheduling problems characterized by various machine capacities and workload distributions. Moreover, analytical bounds will be established to evaluate the objection function and simulation optimization algorithms will be incorporated to further extend the applicability of the developed methods. The new perspective this research brings to the stochastic modeling of the order scheduling problem intends to suggest approached to enhance the effectiveness of various managerial options for improving production efficiency and customer satisfaction in manufacturing systems.

本项目通过随机优化的研究途径，对现代制造行业中常见的"随机客户订单调度"问题进行数学建模和优化。与传统的"确定性"订单调度问题不同，本问题中的订单到达、产品种类、各类产品工作量均为随机，且需对调度进行动态决策。本研究项目拟通过对现场异质加工资源的匹配整合，对具有随机性的成套订单进行生产调度，以期快速响应客户需求。在性质层面，本项目拟通过对目标函数的分解及解空间的分析，探索最优解性质，构建用于系统性能评估的最优解下界。在算法层面，本项目一方面拟通过利用最优解特征，设计高效可靠的动态启发式算法；另一方面拟结合仿真优化理论，构建用于随机调度领域的仿真优化方法体系，实现从专属调度算法向普适优化方法的过渡。本项目拟以订单系统的"随机性"与决策过程的"动态性"为关注点，弥补现有文献研究中的不足，完善订单调度理论，并有效推进其在生产实践中的应用。

本项目通过随机优化的方法，对现代制造行业中常见的“随机客户订单调度”问题进行数学建模和优化。与传统研究中的“确定性”订单调度问题不同，本问题中的订单到.达、产品种类、各类产品工作量均为随机，且需对调度进行动态决策。本项目重点研究了制造现场异质加工环境，针对具有随机性的客户订单进行生产调度，增加吞吐量，降低订单周期，满足客户需求。在数学性质推导方面，本项目通过对目标函数的分解及解空间的分析，给出大量最优解性质，构建用于系统性能评估的最优解下界。在算法涉及层面，本项目通过利用最优解特征，设计了多套高效可靠的动态启发式算法；另一方面拟结合仿真优化理论，对随机调度领域的仿真优化的方法论进行了完善，在理论上建立了从专属调度算法向普适优化方法的过渡桥梁。本项目在订单环境“随机性”与决策过程的“动态性”方面弥补了现有文献研究中的不足，完善订单调度理论，并在汽车维修、营养品生产领域实现了理论的落地应用。