混装生产精准作业的多尺度智能元胞调度模型研究

项目名称： 混装生产精准作业的多尺度智能元胞调度模型研究

项目编号： No.50875190

项目类型： 面上项目

立项/批准年度： 2009

项目学科： 金属学与金属工艺

项目作者： 唐秋华

作者单位： 武汉科技大学

项目金额： 33万元

中文摘要： 生产计划与作业调度的分离、大规模系统计算的延迟、调度性能检验的滞后，是混合装配生产管理的三大软障碍，有效的解决策略是混装生产精准作业调度。该调度方法是多参数、多阶段、多变量和多目标的高维组合优化，计算复杂度为NP-hard，研制其高效快速求解算法具有显著价值。 同步重构任务分配和投产序列，可以集成生产计划和作业调度，同时实现生产高效和作业精准的目标，将多产品混合装配生产管理上升到精准作业调度的层次。本项目以混装生产线为研究对象，通过解析工位构造、时间、产品类型，实现了具有多时间尺度的生产计划和作业控制的集成调度系统框架，该集成框架的多尺度分解机理保证了其对市场需求和作业控制的有效反馈。在遗传、蚁群等算法基础上，分析元胞与邻域间微观作用，推理元胞进化规则，通过元胞自动机的演变使整个系统达到平衡态。 各元胞的同步并行计算，保证了多产品任务分配的运算速度；元胞自动机和多Agent系统相结合，加速了多尺度非线性耦合调度系统的运算；智能元胞算法的微观作用机制保证了大规模计算的有效性。本项目研究有益于提高多产品混装调度作业性能，提升管理绩效，促进管理体系、技术和算法的发展。

中文关键词： 集成生产计划与调度；元胞自动机；多尺度分解；多Agent系统；

英文摘要： Three barriers have been impeding the production management of mixed-model assembly lines: separation of production planning from operation scheduling, exhausted computational effort for large scale system, time delay of performance verification. Its solution, precise operation scheduling of mixed-model assembly lines, is a high dimensional combinatorial optimization problem with multiple parameters, stages, variables and objectives with NP-hard computational complexity. Developing efficient algorithms to solve such engineering problems is significant. By complementing task allocation and production sequencing synchronously, the production planning and operation scheduling are integrated, two objectives of production efficiency and operation precision can be achieved, and the production management of mixed-model assembly lines can be promoted to the level of precise operation scheduling. Focusing on mixed-model assembly lines, this work proposed an integrated framework of production planning and operation control with different scales in station complexity, time period and product variety, which guaranteed the feedback to market requirement and operation control efficiently and effectively. Then based on genetic algorithm, ant colony algorithm, the interaction between each cellular and its neighborhood was analyzed and formulated into evolution rules of cellular space. Under the mechanism of intelligent cellular automata, workloads among stations are balanced ultimately. The parallel computation in cellular automata improved the solution of multiple product task allocation, the hybrid algorithm which combined cellular automata and multi-agent system speeded up the computation of multi-scale nonlinear coupling scheduling system, and the mechanism of intelligent cellular algorithm ensured the validity of large scale computation. This work is helpful for promoting the operation performance on mixed-model assembly lines, improving production scheduling and enhancing the development in management system, technology, approach and algorithms.

英文关键词： Integrated production planning and scheduling; Celluar automata; Multi-scale decomposition; Multi-agent system;