基于学习—遗忘曲线的多周期人机单元构建研究

发布时间：2018-05-06 09:42

本文选题：单元构建 + 多周期　；参考：《浙江工业大学》2012年硕士论文

【摘要】：单元制造系统因为能兼顾大批量制造的经济性和中小批量的灵活性,而受到学术界和企业的广泛关注,单元构建作为单元制造系统设计的首要问题,也成为关注的重点。本文通过对单元构建技术和学习-遗忘曲线相关理论的研究,考虑人的学习-遗忘效应,在多个周期下建立单元构建的数学模型。本文的研究内容主要包括：首先,将学习-遗忘曲线与单元构建问题相结合,建立在多个周期下的单元构建数学模型,以单元间物料处理时间、机床载荷偏差及机床操作人员加工准备时间为目标函数,引入数值算例,验证模型的有效性。其次,对于数值算例,利用遗传算法的相关设计流程,对模型进行求解,并得到相应的最优单元划分方案和人员分配方案。在遗传算法的设计方面,选用“轮盘赌”和“精英保留”策略对个体进行选择,对于人员信息位的变异部分,采取“关联变异”和惩罚函数相结合的方法,保证种群的进化同时排除非可行解。再次,结合工程实例,根据公司的历史数据,利用线性回归法建立各项操作的学习率表,通过市场调查和历史数据建立多个周期零件的需求表,结合基于学习-遗忘曲线的多周期单元构建数学模型,运用遗传算法进行求解,得到最优解下的单元划分方案。
[Abstract]:Cell manufacturing system (CMS) has been widely concerned by academia and enterprises because it can take into account the economy of mass manufacturing and flexibility of medium and small batch manufacturing. Cell construction as the primary issue of cell manufacturing system design has also become the focus of attention. In this paper, the mathematical model of unit construction is established with the consideration of human learning-forgetfulness effect through the study of unit construction technology and the theory of learning-forgetting curve. The main contents of this paper are as follows: First of all, the learning-forgetting curve is combined with the problem of unit construction, and the mathematical model of unit construction is established under multiple cycles. The objective function is the material processing time between units, the load deviation of machine tool and the processing preparation time of machine operators. A numerical example is introduced to verify the validity of the model. Secondly, for numerical examples, the model is solved by using the relevant design flow of genetic algorithm, and the corresponding optimal unit partition scheme and personnel allocation scheme are obtained. In the design of genetic algorithm, the strategy of "roulette" and "elite reservation" is used to select individuals, and the method of "association mutation" and penalty function is adopted for the mutation part of personnel information level. The evolution of the population is guaranteed and the infeasible solution is excluded at the same time. Thirdly, according to the historical data of the company, the learning rate table of each operation is established by linear regression method, and the demand table of multi-cycle parts is established by market survey and historical data. A mathematical model based on the learning-forgetting curve is constructed, and the genetic algorithm is used to solve the problem, and the unit partition scheme under the optimal solution is obtained.
【学位授予单位】：浙江工业大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TH165

【参考文献】