Robust multi-objective job-shop scheduling for make-to-order manufacturing

有效的订单排序和任务调度已成为在当今竞争激烈的市场中生存的必要条件。制造环境中的调度涉及将已发布的订单转换为具有相关截止日期的作业，并将其分配到指定的工作站。车间作业调度是一个多目标组合优化问题。尽管如此，大多数调度算法通常在给定的目标函数下是有效的，而不是它们的组合。此外，车间发生的不可预见的事件，如机器故障以及紧急/取消订单可能会对计划产生重大影响。为了减少对初始计划进行重大更改的需要，应该将鲁棒性标准纳入计划模型中。在这个项目中，申请人将与Festo教学有限公司合作，以克服公司最近在安排不断增长的销售方面所面临的挑战。主要目标是开发一个鲁棒多目标作业车间调度算法程序库。实现这一目标的第一步是通过调查与不同产品族相对应的制造过程来描述具体的调度问题。除了最相关的干扰来源外，确定具有相应优先级的各种决策标准也将是项目中的重要步骤。新的调度引擎预计将集成到他们当前的生产计划软件中，将包含一组基本和复合调度规则，以及本地搜索和元启发式算法，以生成替代调度。开发一个离散事件模拟模型，以评估生成的时间表的性能，也将在本研究项目的范围内。该项目的结果将为公司在制造交货期和在制品库存减少方面提供重要的好处。此外，还可以集成已开发的调度引擎

Effective order sequencing and task scheduling has become a necessity for survival in today's competitivemarkets. Scheduling in a manufacturing environment concerns with translating released orders into jobs withassociated due dates and allocating them to designated workstations. Job-shop scheduling is a multi-objectivecombinatorial optimization problem. Nonetheless, the majority of scheduling algorithms are usually efficientunder a given objective function rather than a combination of them. Furthermore, unforeseen events on theshop floor, such as machine break-downs along with rush/canceled orders may have a significant impact on theschedule. In order to reduce the need for major changes in the initial schedule, robustness criteria should beincorporated into scheduling models.In this project, the applicant will team up with Festo Didactic Ltd. in order to overcome the challenges thecompany has been recently confronted in scheduling the continuously increasing sales. The main goal is todevelop a library of algorithmic procedures for robust multi-objective job-shop scheduling. The first steptowards achieving this goal would be the characterization of the specific scheduling problem via investigationof manufacturing processes corresponding to different product families. Identification of variousdecision-making criteria with corresponding priorities in addition to the most pertinent sources of disruptionswill be also an essential step in the project. The new scheduling engine that is expected to be integrated intotheir current production planning software will incorporate a set of basic and composite dispatching rules alongwith local search and metaheuristic algorithms to generate alternative schedules. Developing a discrete-eventsimulation model for the purpose of assessing the performance of generated schedules would be also within therealm of this research project.The results of this project will provide essential benefits to the company in terms of manufacturing lead-timeand work-in-process inventory reduction. Furthermore, the developed scheduling engine can be incorporated