面向低熵的一类多态性作业车间布局智能建模与稳健协同优化

According to the development model of low entropy and the pressing needs of the development of domestic original equipment manufacturing enterprises, a new problem of robust collaborative optimization of a type of polymorphism job shop layout for low entropy has been proposed. This research project described prototype characteristics, prototype parts and details of polymorphism job shop layout based on Boolean parameters division, formulated low entropy constraints which combines cellular layout and logistics paths design, built a threelevel, two dimension swarm intelligent cellular automata topological structure which describes the physical layer, exercise layer and disturbance layer of shop layout, defined property of a sixtuple multi-population cellular automata and mutative spatial scale, abstracted a self-evolution mechanism of cellular automata, and optimized evolution rules by using genetic algorithm. Solve the dynamic problems of layout by intrduceing cellular model space to multi-intelligent mechanism and using packaging and expanding to make cellular automata in exercise layer and disturbance layer intelligent cellular automata which has autonomy. Then, the project built a multi-target indicator function for layout, made physical entropy measure of robust analysis of shop layout based on key indicator collaborative analysis of job shop scheduling. At last, built an optimization model of polymorphism job shop layout that can reflect the practical productive circumstance, developed a laboratory simulation and empirical research of typical enterprise. The proposed new approach of layout modeling and optimization provides effective analytical methods and tools a robust collaborative optimization of a type of complex job shop layout with operational targets based on low entropy, so it has important realistic significance and theoretical significance.

依据低熵化发展模式和国内原始设备制造企业发展的迫切需求，提出面向低熵的一类多态性作业车间布局稳健协同优化新问题。基于布尔化参数划分描述多态性作业车间布局原型特征、原型局部和细节，表征集单元布局和物流路径设计为一体的低熵约束，构建描述车间布局实体层、作业层及扰动层的三层次二维多种群元胞拓扑结构，定义六元组多种群元胞属性和可变元胞空间，抽象元胞自演化机制，采用遗传算法优化演化规则。将多智能体机理引入元胞空间，封装并扩展作业层和扰动层元胞为具有自主性的智能元胞，解决布局动态性问题。建立布局多目标指标函数，在车间调度关键指标协同分析基础上开展以物理熵为度量的车间布局稳健性分析。最终构建反映生产实际的多态性作业车间布局优化模型，开展实验室模拟和典型企业实证研究。项目提出的布局建模和优化新方法为低熵运行目标下一类复杂作业车间布局稳健协同优化提供可行有效的方法与工具，具有非常重要的现实意义和理论意义。

依据低熵化发展模式和国内原始设备制造企业发展的迫切需求，提出面向低熵的一类多态性作业车间布局稳健协同优化新问题。分析面向低熵多态性作业车间布局的原型特征，重点描述加工和转运批量动态化、设备机台多样性和动态性、不同产品非线性加工路线差异、在制品管控、物料搬运设备限制、物流路径变化、订单扰动等的原型局部和局部细节。在多态性作业车间布局原型特征基础上表征集单元布局和物流路径设计为一体的低熵约束，分析设备排列顺序、物流路径容量限制、网络构建费用和物料搬运费用间平衡、局部不规则的布局空间、单元间路径不能穿越制造资源等问题。构建车间布局的三层次二维群元胞拓扑结构，设备、人员、物料表征实体层，加工、装配、搬运、装卸、储存等设定为作业层，订单管理、工艺排序、设备管理、人员管理等设定为扰动层，广义制造单元设定为实体层格点，规划实体元胞的可变元胞空间，元胞属性映射某一种群的品种、工艺、制造单元等特征向量。根据原型特征与集成布局动态约束的格点属性设定和元胞自演化机制抽象，基于禁忌搜索方法获取规则决策元初解，在初始自演化规则建立的基础上进行基于遗传算法的演化规则优化。开展群智能元胞机理研究，定义、封装作业层和扰动层元胞，扩展为具有自主性的智能型元胞，解决布局的动态性问题。基于企业实际需求，建立面对低熵模式的包含物流量距、设备成本、车间面积利用率、在制品库存量、布局柔性车间布局等的多目标指标函数，结合车间调度关键指标协同分析开展以物理熵为度量的车间布局稳健性抽象和描述。构建和求解反映生产实际的多态性作业车间布局稳健协同优化模型，基于SWARM平台开发原型系统。开展实验室模拟和原型系统典型企业实证分析，验证模型和算法的正确性，同时进一步完善模型。项目提出的布局建模和优化新方法为低熵运行目标下一类复杂作业车间布局稳健协同优化提供可行有效的方法与工具，具有非常重要的现实意义和理论意义。