基于多色集合理论的工艺规划和生产调度集成建模理论与优化方法

For the middle and small-sized Supporting manufacturing enterprises, to cope with the complicated and dynamic environments with varied demands and limited resources, integrating process planning and scheduling (IPPS) to make scientific decisions is a key problem needed to resolve in their manufacturing process, thus, this project proposes to research the integrated modeling and optimization methods for the manufacturing enterprises’ IPPS problems under the dynamic environments. Specifically, the research contents have three points: (1) By utilizing the polychromatic set model with unified form to describe the constraint relationships and dynamic characteristics in the process planning and scheduling, and to express the IPPS model, this project establishes computational models of the complicated multiobjective IPPS problems. (2) Based on the property that the distribution structures of the solutions of multiobjective optimization problems present regularity, the manifold learning algorithms are adopted to identify the structure information of the multiobjective IPPS problems’ solutions. This information is used to direct the design of the algorithms, and it constructs a fast solving hybrid algorithm framework based on the computational model. (3) Based on the theoretical research, a IPPS prototype system of the middle and small-sized PCB manufacturing enterprises is developed to implement, verify and improve the research achievements. The theory and methods proposed in this project will be able to provide new thoughts and techniques for the middle and small-sized manufacturing enterprises to improve their producing efficiency and response capability to the market change.

处于供应链中的中小制造企业如何应对需求多变、资源制约的复杂动态环境是其在市场竞争中急需解决的关键问题，基于此，本项目提出基于多色集合理论研究动态环境下制造企业工艺规划和生产调度集成（IPPS）建模与优化方法。具体内容为三点：（1）利用形式统一的多色集合模型实现工艺规划与生产调度中约束关系、动态特性的描述以及IPPS模型的表达，建立复杂多目标IPPS问题的可计算模型；（2）基于多目标问题最优解分布结构具有规则性的性质，运用流形学习识别多目标IPPS问题解的结构信息，用以指导算法设计，构建出基于可计算模型的快速求解混合优化算法框架；（3）在理论研究基础上，开发印刷电路板（PCB）制造企业工艺规划与生产调度集成优化原型系统，对研究成果进行应用、验证与完善。本项目所研究的理论和方法能够为中小制造企业改进生产效率，提高应对多变市场的快速响应能力提供新的思路和技术。

处于供应链中的中小制造企业如何应对需求多变、资源制约的复杂动态环境是其在市场竞争中急需解决的关键问题，基于此，本项目提出了基于多色集合理论研究动态环境下制造企业工艺规划和生产调度集成（IPPS）建模与优化方法。. 主要研究工作及成果为：1.建立了柔性车间调度数学模型和基于多色集合的统一可计算约束模型。针对求解柔性车间调度中所存在的早熟或收敛过慢等不足，提出了基于多色集合理论的元胞遗传算法。多色集合能够运用结构相对简单的围道矩阵建立具有标准形式的递阶结构模型，实现复杂优化问题的描述。因此运用了围道矩阵柔性车间调度中的设备约束和工艺约束进行了描述，通过搜索围道矩阵，缩小解的空间从而提高了求解效率；2.通过经典实例对比测试，验证了改进元胞多目标遗传算法在收敛性、分布性及覆盖性方面均有良好表现。基于柔性车间调度为多目标优化问题，在选择操作中，根据邻居中个体的适应度来自适应选择交叉个体，有效地防止种群过早收敛。元胞遗传算法中相邻个体的邻居互相重叠，为算法提供了隐性迁移机制且保持了种群多样性，从而降低了算法的时间和空间复杂度，有利于求解多目标优化问题；3.将约束模型与元胞多目标遗传算法的编码、种群初始化、遗传操作、适应度计算等环节相结合，构成了混合优化算法体系。运用计算机仿真，验证了所提出方法在求解柔性车间调度问题上的合理性和高效性；4.基于多色集合约束模型的柔性车间调度系统的总体设计和实现。对柔性车间调度系统总体框架和模块进行规划分析，对分功能模块进行设计实现，开发了车间调度原型系统。. 本项目所研究的理论和方法能够为中小制造企业改进生产效率，提高应对多变市场的快速响应能力提供新的思路和技术。