纸币清分机产品族设计过程配置优化方法研究

发布时间：2018-06-01 01:38

本文选题：纸币清分机 + 产品族　；参考：《大连理工大学》2013年博士论文

【摘要】：在日益激烈的国际市场竞争环境中,金融机具行业企业面临着尽快提供适合不同层次和群体的产品设计问题。纸币清分机是金融领域现金处理业务的高端必备设备,具有结构复杂、智能化程度高、市场需求变化快的特点。本文针对面向大批量定制的纸币清分机产品族设计过程中一些关键问题进行了深入研究,目的是以最快的速度和最低的成本为客户设计出满足个性化需求的产品,结合纸币清分机开发过程特点,应用产品族设计理论,详细阐述了纸币清分机产品族设计开发过程,并提出了产品族设计过程的配置优化方法,并对研究成果进行了验证和应用。本文的主要研究内容叙述如下： (1)应用基于设计公理的产品族设计理论提出纸币清分机产品族设计过程。建立产品开发过程的多视图模型和任务结构模型,实现任务结构模型到Petri网模型的映射定义并进一步推导与工作流网的映射转换关系,结合纸币清分机的设计开发过程完成了任务结构模型到工作流网的转化并进行了实例验证分析。 (2)提出一种面向产品族设计的纸币清分机产品平台核心模块的确定方法。通过建立纸币清分机产品的模块关联矩阵,采用变个体长度的混合蛙跳算法优化模糊聚类数和聚类中心,求得产品平台构成模块的最优模糊划分,并应用切断算子和拼接算子对个体进行重组形成新个体并进行局部寻优,解决模糊C均值聚类局部极小值和初始化选值敏感性问题。在纸币清分机产品族平台设计中进行了仿真验证,实现纸币清分机模块化产品族公共平台和核心模块的有效确定。 (3)提出了产品族设计过程中产品族配置性能预测方法。在基于递阶支持向量机的配置综合性能预测方法框架下对纸币清分机产品族的历史数据库进行数据挖掘,预测新配置产品的性能,提出一种改进混合蛙跳算法来优化SVM模型的核函数参数和误差惩罚因子,应用模拟退火算法提高算法的局部搜索能力和收敛速度,通过以纸币清分机产品族为实例的仿真实验并对比其他方法验证所提预测方法的有效性,为准确、快速地进行纸币清分机产品族配置提供了有效方法,满足多样化的客户需求。
[Abstract]:In the increasingly fierce international market competition environment, financial machinery industry enterprises are faced with the problem of providing product design suitable for different levels and groups as soon as possible. Banknote sorting machine is the high-end necessary equipment for cash processing in the financial field. It has the characteristics of complex structure, high intelligence and fast changing market demand. In this paper, some key problems in the product family design of the paper money sorter for mass customization are studied in depth. The purpose of this paper is to design the products to meet the individual requirements with the fastest speed and the lowest cost for the customers. Based on the characteristics of the development process of the paper currency sorting machine and the application of the product family design theory, the design and development process of the product family of the paper currency sorting machine is expounded in detail, and the configuration optimization method of the product family design process is put forward. The research results are verified and applied. The main contents of this paper are as follows: 1) applying the product family design theory based on design axiom, the paper currency sorting machine product family design process is put forward. The multi-view model and the task structure model of the product development process are established, and the mapping definition from the task structure model to the Petri net model is realized, and the mapping transformation relationship with the workflow net is further deduced. Combined with the design and development process of banknote sorting machine, the transformation of task structure model to workflow net is completed, and an example is given to verify and analyze it. This paper presents a method for determining the core module of the product platform of the banknote separator designed for the product family. By establishing the module correlation matrix of the paper currency sorter, the fuzzy clustering number and the cluster center are optimized by using the mixed leapfrog algorithm with variable individual length, and the optimal fuzzy partition of the product platform is obtained. Then, cutting operator and splicing operator are used to recombine the individual to form a new individual and to search for local optimization. The problem of sensitivity of fuzzy C-means clustering local minimum and initial selection value is solved. In the design of the product family platform of the paper currency sorting machine, the simulation is carried out to realize the effective determination of the common platform and the core module of the product family of the paper currency sorting machine. 3) the performance prediction method of product family configuration in the process of product family design is put forward. In the framework of hierarchical support vector machine (SVM) based configuration comprehensive performance prediction method, the historical database of the product family of banknote sorting machine is mined to predict the performance of the new configuration product. An improved hybrid leapfrog algorithm is proposed to optimize the kernel function parameters and error penalty factors of SVM model. The simulated annealing algorithm is applied to improve the local search ability and convergence speed of the algorithm. Through the simulation experiment of the product family of banknote sorting machine and comparing with other methods to verify the effectiveness of the proposed prediction method, this paper provides an effective method for accurate and rapid configuration of the product family of the paper currency sorter, and meets the diverse customer needs.
【学位授予单位】：大连理工大学
【学位级别】：博士
【学位授予年份】：2013
【分类号】：TH693.5

【参考文献】