机械加工系统基于实时信息反馈的稳健性调度优化理论、算法与应用

Most studies of traditional static scheduling ignored the dynamic features of production requirements, as well as robustness and stability of production processing, which lead to the gap between theories and applications on scheduling issues. To solve this problem, this project mainly focuses on optimized scheduling theories, algorithms and applications on the stability and robustness of mechanical manufacturing systems based on real-time status feedbacks under dynamic and uncertain circumstances. Firstly, together with Proactive-reactive Mechanism/Mode based on real-time status feedbacks, a dynamic analysis model of machining system based on the Hybrid Stochastic Petri Nets is proposed by classification and description of the dynamic disturbances; Secondly, scheduling algorithms with high efficiency based on Domain Knowledge such as the neighborhood structures and the fitness-landscape, together with the optimization of multi-objective scheduling and its decision-making strategies grounded on the Pareto Domination approach, the Fuzzy Preference and the Fuzzy Integral methods are investigated. Thereafter, the stability and the robustness of scheduling on the basis on of Proactive Mode under uncertain circumstances are probed in by adopting levels of improved β-distributions and Monte Carlo simulations. Grounded on the intelligent re-scheduling decision-making mechanism of an expert system, the multi-objective re-scheduling optimization strategy with time and space is put forward. In addition, an advanced intelligent re-scheduling APS (Advanced Planning and Scheduling) system is developed. Ultimately, an experimental platform integrated hardware with software grounded on real-time status feedbacks is established. The achievements of this project with great significance can provide new theories and methods on the scheduling optimization of machining systems, which can promote the integration of scheduling theories and practices further.

传统静态调度的研究忽略了生产需求的动态性和生产过程的健壮性和稳定性，造成理论研究与应用存在较大差距。本项目围绕动态不确定环境下机械加工系统基于实时信息反馈的稳健性调度理论与方法展开研究。首先，通过扰动分类与描述，结合基于实时信息反馈的主动反应模式，建立基于混合随机Petri网的机械加工系统动态分析模型；其次，对基于邻域结构和适应度地形等领域知识的高效调度优化算法，以及基于Pareto支配、模糊偏好和模糊积分的多目标调度优化和决策进行研究；然后，采用改进β分布、蒙特卡洛仿真法研究不确定环境下基于主动模式的健壮性调度，基于专家系统的智能重调度决策机制，提出面向时间与空间的多目标反应调度策略，开发智能重调度APS高级计划与排程系统；最后，研发基于实时信息反馈的调度优化软硬件一体化平台。本项目将为机械加工系统调度优化提供新的理论与方法，促进调度理论与实际应用相结合，具有重要的理论价值和应用价值。

传统静态调度的研究忽略了生产需求的动态性和生产过程的健壮性和稳定性，造成理论研究与应用存在较大差距。本项目围绕动态不确定环境下机械加工系统基于实时信息反馈的稳健性调度理论与方法展开研究。首先提出多层次建模与智能调度算法相结合的方法，建立了基于智能调度算法的生产运作仿真优化和运行控制系统；随后，深入系统地研究调度问题领域知识，根据问题本质特征，结合群体智能算法和局部搜索各自优势，设计适应问题本质的混合算法，使集中搜索和分散搜索之间达到更合理的平衡，较好解决了调度计算规模大、解空间复杂的难题；然后，基于周期与事件驱动的再调度策略，研究不同的再调度周期下，先后到达工件进行调度得到的最后的完工时间、总拖期、总效率和总稳定性之间的差异，得出了不同再调度周期对各个性能指标的影响，以指导生产实践。最后，基于专家系统的智能重调度决策机制，提出面向时间与空间的多目标反应调度策略，开发智能调度APS高级计划与排程系统，研发基于实时信息反馈的调度优化软硬件一体化平台。本项目研究成果可为机械加工系统调度优化提供新的理论与方法，促进调度理论与实际应用结合，对提升我国制造企业竞争具有重要意义。