【作者】 胡咏梅；

【导师】 贾磊；

【作者基本信息】 山东大学 ， 控制科学与控制工程， 2005， 博士

【摘要】 本文研究了粗集理论(rough sets)和S-粗集(singular rough sets)理论在生产车间动态调度中的应用,在生产车间动态加工环境下,当生产设备损坏与修复、急加工工件到来、新加工工件到来和工件到期时间改变等实时事件发生时,静态调度不能适应动态加工环境的变化,必须采用动态调度对实时事件进行处理,以适应动态的加工环境。本论文基于粗集和S-粗集理论,并与数学规划和调度专家经验相结合,对动态调度窗口中调度工件的再识别和再调度问题,提出了相应的车间动态调度窗口工件识别方法,建立了基于上述理论的车间动态调度模型,并基于遗传算法求解所建立的动态调度模型。与传统动态调度方法相比较,仿真结果验证了所提方法的有效性和正确性,并在减少再调度次数和提高动态调度的稳定性上有一定的优势,且能获得满意的调度结果。 本文的主要工作和创新点如下: 1、阐述了车间调度的概念和意义,建立了一般的车间静态和动态调度模型。为适应车间工件调度的特点和车间动态调度实时性的要求,对遗传算法的遗传操作算子进行了改变,使其交叉算子更适合于车间工件调度模型的求解。仿真实验结果表明了所提算法的有效性和正确性。 2、研究了粗集理论在车间动态调度建模中的应用。基于粗集的分类特性、属性值和近似特性,提出了基于粗集的车间动态粗调度窗口工件识别方法,建立了车间动态粗调度模型。通过仿真实验,验证了所提动态粗调度模型的正确性和有效性。与传统动态调度方法相比,该方法减少了动态再调度的次数,均衡了设备的利用率,而且能够获得满意的调度结果。 3、研究了遗传算法在车间动态粗调度中的应用。给出了基于遗传算法山东大学博士学位论文的车间动态粗调度算法,仿真实验表明了该算法的有效性。4、研究了S一粗集理论在车间动态调度建模中的应用,把调度专家经验有效地应用于模型参数设定和S一粗集相关属性值设定上。基于S一粗集特有的副集、动态近似特性和元素迁移特性,提出了基于S一粗集的车间动态S一粗调度窗口工件识别方法,建立了基于S一粗集的车间动态S-粗调度模型。与动态粗调度模型相比,动态S一粗调度模型不但具有动态粗调度模型的优点,而且,能够更好地描述车间动态调度的动态特性,实施起来更有理论依据、更系统、更方便。仿真实验结果展示,与动态粗调度方法相比较,因采用了相同等价类工件全部加工完成启动一次再调度的策略,能进一步减少动态再调度次数,并获得满意的调度结果。5、研究了遗传算法在车间动态S一粗调度中的应用,给出了基于遗传算法的车间动态S一粗调度算法,仿真结果验证了该算法的有效性和优越性。 最后,总结了本文的主要工作,指出了进一步的研究方向。关键词:动态调度,动态识别,粗集,S一粗集,车间调度 本文的研究工作得到国家863高科技项目的资助。更多还原

【Abstract】 This dissertation studies the application of rough sets theory and S-rough sets ( singular rough sets ) theory in job shop dynamic scheduling. We study the job re-identification and job re-scheduling problems in the job shop dynamic scheduling window under dynamic processing environment. The corresponding job identification methods of the job shop dynamic scheduling window are presented, in the case of the machine failure and repair, the due data change of jobs and the urgent jobs coming. The new methods are the organic combinations of rough sets or S-rough sets, the mathematics programming and the scheduling expert experience. We establish the job shop dynamic scheduling modeling based on these theories, and resolve these models using an effective genetic algorithm. The simulation results demonstrate the novel methods compared with the traditional dynamic scheduling methods have certain merits in the decrease of the re-scheduling degree and the improvement of the system stability.The main study works of this dissertation are as follows.1. At first, we give the concept and significance of job shop scheduling, and establish the general job shop static and dynamic scheduling models. The algorithm cross operators of the genetic algorithm is changed for the characteristic of job shop scheduling and the real time demand of job shop dynamic scheduling. The simulation results show the availability and correctness of the proposed algorithm.2. Secondly, as a novel application of rough sets theory, the job shop dynamic scheduling modeling is studied. Based on the approximate characteristic of rough sets, a job identification method of the job shop dynamic rough scheduling window is presented, the correlation conceptions of job shop dynamic rough scheduling are defined, and the job shop dynamic rough scheduling model based on rough sets is established. The simulation results demonstrate the correctness and the advantage of the proposed methods compared with the traditional dynamic scheduling methods. Through the use of this dynamic scheduling algorithm, firstly the dynamic processing environment is adapted and a satisfied scheduling result is obtained, secondly, the problem dimension is reduced and the utility ratio of the equipment is advanced, finally the rescheduling degree is decreased and the system stability is improved.3. Thirdly, we study the application of genetic algorithm in the jobshop dynamic rough scheduling, and present a dynamic rough scheduling algorithm based on genetic algorithm. The simulation results show the availability of the proposed algorithm.4. As follow, as another novel application of S-rough sets theory based on its assistant set, its dynamic approximate characteristic and its element transfer characteristic, the job shop dynamic scheduling modeling is studied. The scheduling expert experiences are effectively applied to the model parameter setting and the relative attribute value setting of rough sets. A job identification method of the job shop dynamic S-rough scheduling window is presented, the correlation conceptions of job shop dynamic S-rough scheduling are defined, and the job shop dynamic S-rough scheduling model based on S-rough sets is established. Not only dynamic S-rough scheduling model has the merits of dynamic rough scheduling model, but also it better describes the dynamic characteristic of job shop dynamic scheduling, and its implement is more convenience, more system, and has more theory basis compared with the dynamic rough scheduling model. The simulation results show the correctness and the advantage of the proposed methods. By the use of the dynamic S-rough scheduling algorithm, the rescheduling degree is decreased farther, and the satisfied scheduling result is obtained,5. And then, we study the application of genetic algorithm in the job shop dynamic S-rough scheduling, and present a dynamic S-rough scheduling algorithm based on genetic algorithm. The simulation results show the availability更多还原