多品种多批次混合Flow-shop车间的复合规则类调度算法及其理论分析

Considering the characteristics of hybrid flow shop with multiple types and multiple batches, this project deals with the hybrid flow shop scheduling problem of energy optimization subjected to multiple kinds of energy-consumed media and coupling effect of multiple groups of decision variables. The content includes: building up decomposition mathematical models based on multiple kinds of energy-consumed media and coupling effect of multiple groups of decision variables; regarding to these models, a frame of generating algorithms based on the combination of heuristics. By embedding different heuristic rules in different nodes of the frame, lots of combinable algorithms are produced; from algorithm theories, the performances of these combinational rules are tested by analysis of their time complexities, worst-case rates and average rate; finally, simulated experiments are designed to test these mathematical models and combinational rules. Creativities of this project lie in the following: regarding of the characteristics of the problem, decomposition models and a frame of algorithms based on combinational heurists are proposed; a mathematics analysis method based on structure is proposed to obtain the worst-case rate of the combinational rules; a probabilistic method proposed to analyze the average rate of the combinational rules. The output of this project can provide the manufacturers which is order-driven, massive and high energy-consumed with theoretical basis to make the production plan.

针对多品种多批次的混合流水车间生产特点，本项目对具有多种能源介质和多组决策变量耦合的混合flow-shop车间调度问题展开了研究。具体包括：建立具有多种能源介质多组决策变量耦合效应分解模型；基于分解模型，提出一种基于启发式规则的复合算法生成框架，通过在算法框架中的不同节点嵌入不同的启发式规则，可以产生大量不同的复合规则算法；通过对算法的时间复杂度分析、最坏情形界分析和平均效果分析，从算法理论角度评价了所生成复合规则类算法的优劣性；最后通过设计算法仿真实验对上述模型和算法进行验证。特色与创新之处体现在：构建了一种基于问题特征的分解模型，并提出一种基于启发式规则的复合算法框架；提出了一种构造性的数学分析方法，获得了复合规则类算法的最坏情形界；提出了一种概率方法，用于分析复合规则类算法的平均效果。研究成果将为大批量定制化高耗能企业制定生产计划时，提供了复合规则类算法的构建、选择与评价的理论基础。

针对多品种多批次的混合流水车间生产特点，本项目对具有多种能源介质和多组决策变量耦合的混合flow-shop车间调度问题展开了研究，旨在为大批量定制化高耗能企业制定生产计划时，提供了复合规则类算法的构建、选择与评价的理论基础。围绕上述目标的实现，并结合课题组前期对企业的调研，对以下内容展开了研究：. 1. 建立具有多能源介质多组决策变量耦合效应的分解模型。分别针对陶瓷行业两大高耗能设备（抛光机和隧道窑）为研究对象，分析其能耗的组成机理，建立其能耗参数模型，并提出相应的优化算法对设备参数进行优化使得设备运行能耗最少。. 2. 构建基于启发式规则的复合算法框架及其算法的评价和分析。分别针对混合柔性流水车间、智能车间等不同类型车间构建其复合算法框架，并对复合规则算法的时间复杂度和平均求解效果进行了验证。. 3. 算法插件的开发与企业应用验证。课题组基于项目研究成果，针对一家生产3C产品的企业设计开发了一套多品种中小零件智能车间MES调度系统，为企业实现了自适应排产功能，并在企业中得到初步应用。