面向模具生产的工艺与车间调度紧耦合集成规划

Integrated process planning and job shop scheduling (IPPS) is a potential mechanism to improve the whole efficiency of manufacturing systems. The mould-oriented IPPS will be established based on the structure property of mould manufacturing process planning and scheduling. However, current theory and methods for IPPS cannot completely meet the special requirement of mould manufacturing. The theory investigation and practice exploration for the key scientific problem of mould-oriented IPPS will be carried out in this research. An integrated constraint and optimization model with feature as its core based on Object -Oriented, Polychromatic Sets (PS) and Mixed Integer Programming modeling theory will be established; and an analytical mechanism for the generation of feasible integrated plan will be designed and embedded in static/dynamic optimization algorithm. On this basis, a hybrid method by combining Genetic Algorithm with Tabu Search will be developed by introducing the strategies of inferior parents selection and population degeneration; meanwhile, a hybrid multi-objective optimization algorithm for IPPS based on Pareto-based scale-independent fitness function and vector evaluated method will be proposed; dynamic optimization strategies for different situations will also be provided with adjusted integration model and analytical mechanism. Finally, the above proposed mechanisms will be verified by an integrated instance with many mould parts. As a result, process and scheduling plan will be optimized and determined collaboratively, and the confliction between different optimization objectives can be reduced; meanwhile, the dynamic situations can be dealt with quick response and the whole efficiency of manufacturing system can be improved. This research facilitates the understanding of mould production planning problem and broadens new planning attempt for the problem; and it also provides innovative systematic mechanism for the development of integrated model and static/ dynamic optimization of IPPS.

工艺与车间调度集成规划是整体上提高制造系统效率的潜在机制。本项目在深入研究模具工艺规划和车间调度问题结构特性基础上，提出了面向模具生产的集成规划新模式。针对现有相关理论方法不足和难以满足模具生产集成规划的特殊需求，基于对象化、多色集和混合整数规划建模理论，建立以特征对象为核心的集成规划约束与优化模型；设计易融入集成算法和扩展到动态场景的可行集成方案解析生成机制；基于保劣性选择父个体和衰退种群策略开发遗传算法与禁忌搜索结合的复合机制；引入Pareto独立标量适应度函数和向量评估方法开发多目标优化复合算法；调整集成规划模型和解析机制，开发针对不同动态场景的优化策略并以部分模具零件进行验证。目标是协同优化确定模具零件工艺与调度方案，消减目标冲突，快速响应动态场景，整体上提高制造系统效率。本研究对于理解模具生产规划问题和拓宽求解思路是一种新的尝试；为集成规划建模和静动态优化也提供了系统性创新机制。

背景：开展工艺与车间调度集成(Integrated Process Planning and Job Shop Scheduling, IPPS)是提高模具整体生产效率的潜在机制。但当前面向模具的IPPS在模型构建和优化效果提升两方面仍有待深入研究。基于柔性工艺路线和基于特征的建模是IPPS常见的两类建模方法。基于柔性工艺路线的建模工艺组合度大、部分约束还具有不确定性、模糊性和动态特性，工艺人员难以全面把握约束组合空间内复杂关系，易生成不可行工艺，同时易遗漏可行或排它性地排斥掉部分工艺方案；基于特征的优化机制集成任务进行了分解，不利于全局优化算法设计与全局优化结果的获得。在优化实现方面，需要综合考虑模具加工工时的不确定性、动态特性等，同时有待进一步提升算法效率和效果。.研究内容：本研究集中于集成业务对象和约束描述、优化模型构建、可行集成方案生成、静/动态集成优化以及原型系统开发等。.结果、关键数据和科学意义：研究结合工艺人员经验知识与多色集形式化描述和数理逻辑运算机制分阶段生成柔性工艺网络，为模具车间IPPS优化的约束关系描述与可行工艺方案解析生成提供了新的机制。提出了交叉熵优化机制和激素调节自适应遗传算法。在综合考虑其不确定性基础上，引入了时迁Petri网与启发式A*搜索算法；为避免IPPS状态空间爆炸，同时选择有价值的节点，提出了一种新的动态窗方法用以提高算法效率；获得调度策略后，利用Petri网仿真出系统的动态变化规律。所提出的优化机制较其它对比算法可得到更短的加工周期、平均流通时间和较高的机床利用率；动态调度时，所提出方法在优化加工周期、维持原有方案优化结果上具有明显优势。整体上相应优化机制在提高车间静、动态环境下的柔性、优化分配资源任务、维持工艺与调度两者优化结果（减少目标冲突）等方面具有明显优势，为IPPS特别是模具生产的静动态集成优化实现提供了新机制。最后开发了支持模具IPPS集成的原型系统。