面向作业岛和流水线并存的生产车间调度建模及算法实现

发布时间：2018-04-08 07:43

  本文选题：作业岛　切入点：生产调度　出处：《电子科技大学》2014年硕士论文

【摘要】：本课题是基于企业横向项目“精密空调制造设施布局及生产管理平台的研发”的研究。面向作业岛和流水线共存车间的调度问题是一类重要的调度问题,这种调度是将车间的各个生产装配区域作为研究对象,将订单中要求生产的所有产品合理的分配到各个装配区域(剩余产能的分配),以达到某些目标函数最优的目的;同时也要对多个零件加工单元不同零部件协同排序研究,因为这为装配区域上面产品分配提供必要的物料齐套时间约束。所以,本课题的研究主要分为两个阶段的研究:第一个阶段是面向多个零部件加工单元协同排序问题的研究,可以核算物料齐套时间,为下一阶段的研究提供重要的约束依据;第二阶段,在是否具有物料齐套时间约束的条件下装配区域剩余产能分配问题。对于面向多个零部件加工单元协同排序问题,文章将以最小化N个产品的物料齐套时间加权和作为目标函数建立连续型的数学规划模型,为产品物料齐套时间的综合优化和求解提供理论基础。由于随着问题规模的扩大,数学模型的可行解也会呈现指数增长,一般的求解方法难以实现,鉴于遗传算法具有很好的全局搜素能力而后期迭代不足,对于初始种群的要求较高,所以通过启发式算法迭代出效果相对不错的初始种群,运用遗传算法进行迭代。通过实例的验证,改进的遗传算法可以很好的解决面向多个零部件单元协同排序问题的求解。在对面向多个零部件加工单元协同排序问题的研究的基础之上,对于装配区域剩余产能分配问题进行研究,建立以最小化拖期惩罚费用为目标函数的数学规划模型。首先,在不考虑物料齐套时间约束下建立数学模型,然后在此模型的基础之上,将物料齐套时间约束考虑进调度模型,建立相对完善的模型;其次,对于考虑物料齐套时间约束情况下剩余产能分配模型进行算法设计,也是通过启发式算法对遗传算法的改进进行实现;最后,结合相关的实例对模型及其算法进行验证。最后,本文通过搜集某精密空调制造厂的相关数据,对本文改进的算法和传统的算法,从几个给定的指标进行评估,验证改进的算法比较优越。
[Abstract]:This project is based on the horizontal project of the precision air-conditioning manufacturing facility layout and production management platform research and development.Job-island and pipeline co-existence job-shop scheduling problem is a kind of important scheduling problem, which takes each production and assembly area of the workshop as the research object.In order to achieve the purpose of optimizing some objective functions, all the products required in the order should be reasonably distributed to each assembly area (the distribution of surplus capacity). At the same time, it is also necessary to study the collaborative sorting of different parts in multiple parts processing units.This provides necessary material alignment time constraints for product allocation above the assembly area.Therefore, the research of this topic is divided into two stages: the first stage is the research of collaborative scheduling problem for multiple parts processing units, which can calculate the material alignment time, and provide an important constraint basis for the next stage of research;In the second stage, the problem of distribution of surplus capacity in assembly area is given under the condition of whether or not there is a time constraint on the whole set of materials.In this paper, a continuous mathematical programming model is established based on minimizing the weighted sum of the materials of N products as the objective function to solve the problem of collaborative scheduling for multiple parts processing units.It provides a theoretical basis for the comprehensive optimization and solution of product material alignment time.With the expansion of the scale of the problem, the feasible solution of the mathematical model will increase exponentially, and the general solution method is difficult to be realized. In view of the fact that the genetic algorithm has a good global search ability and the late iteration is insufficient,The requirement of initial population is high, so the heuristic algorithm is used to iterate out the initial population with relatively good effect, and the genetic algorithm is used to iterate the initial population.Through the verification of examples, the improved genetic algorithm can solve the cooperative sorting problem of multiple parts units well.Based on the research of collaborative scheduling problem for multiple parts processing units, the problem of spare capacity allocation in assembly area is studied, and a mathematical programming model is established, which takes minimizing the penalty cost of tardiness as the objective function.First of all, the mathematical model is established without considering the time constraints of the whole set of materials, and then, on the basis of the model, a relatively perfect model is established by considering the time constraints of the whole set of materials into the scheduling model.For the algorithm design of the residual capacity allocation model considering the time constraint of the whole set of materials, the genetic algorithm is improved by heuristic algorithm. Finally, the model and its algorithm are verified with relevant examples.Finally, by collecting the relevant data of a precision air conditioning factory, this paper evaluates the improved algorithm and the traditional algorithm from several given indicators, and verifies that the improved algorithm is superior.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TB497;TP18
，

本文编号：1720687