【作者】 陈响玲；

【导师】 陈雪波； 李梅娟；

【作者基本信息】 辽宁科技大学 ， 控制理论与控制工程， 2012， 硕士

【摘要】 产品数据管理(Product Data Management,PDM)是一种以企业的产品为中心,通过计算机网络和数据库等技术,对所有描述产品的数据和与产品开发相关的过程统一进行管理的技术。近年来针对PDM技术的研究大多数是针对PDM的体系结构、产品的数据模型、软件系统的开发、实施方法和集成接口等,很少有论文研究PDM关键过程中的优化问题,其中基于PDM任务指派问题属于PDM关键过程中优化问题中的一类问题。研究基于PDM任务指派问题的解决方法可为PDM系统核心的设计、开发提供理论和算法的支持,可为提高企业的产品设计和生产效率起到积极的意义。本文用遗传算法的理论和方法对PDM系统项目管理中设计任务的指派问题的建模和优化进行了研究,主要工作分为三个部分:(1)建立了基于PDM任务指派问题的数学模型,目标是确定最优的指派方案,使项目组的有限设计资源得到合理利用。(2)分别从宏观和微观上改进了遗传算法,在宏观上引入了循环策略和模拟退火算法,在微观上针对基于PDM任务指派问题数学模型的特点对编码的方式、解码的规则以及遗传算子进行了改进,提出了一种间接编码的混合遗传算法,并对该混合遗传算法的染色体编码方案、解码规则、适值函数和遗传算子等进行了详细设计。(3)在运用Camel函数和Shubert函数验证了采用循环策略的模拟退火遗传算法性能的基础上,针对提出的间接编码的混合遗传算法,通过一个应用实例,分析并验证了基于PDM任务指派问题的模型和间接编码的混合遗传算法解决该问题的有效性。更多还原

【Abstract】 Product Data Management (PDM) is a kind of technology which concentrates on products of manufacturing enterprises. It manages both the data to describe products and the developing process of products based on computer network and database technology.In recent years there are many literatures relating to research of PDM technology. Most of them concern on the architecture of PDM, models of product data, development of software system, implementation methodology and integration interface, etc. The optimization problems of key processes in PDM are seldom studied, for example, research of the task assignment problem based on PDM. The solution of the task assignment problem in PDM may provide theoretical and algorithmic support for kernel design and development of PDM system. It is meaningful for improving design and manufacturing efficiency in enterprises.In this paper, we discuss modeling and optimization for the design task assignment problem of project management in PDM based on genetic algorithms. The main content includes three aspects as follows:(1) A mathematical model of the task assignment problem in PDM is established, the objective is to obtain the optimal assignment scheme and to enable the efficient allocation of limited design resources. (2)The thesis proposes two ways of macroscopic and microscopic respectively in order to improve the performance of the genetic algorithm. In macroscopic aspect, the circular strategy and simulated annealing algorithm are introduced, and in microscopic aspect, the encoding scheme, decoding rule and genetic operators are improved for the mathematical model of task assignment problem in PDM. Then, we propose a kind of indirect hybrid genetic algorithm. Finally, a kind of encoding scheme, decoding rule, fitness function and genetic operators are designed in detail.(3)Based on using Camel function and the Shubert function to test the performance of the simulated annealing genetic algorithm adopting circular strategy, test was carried through an application example of the task assignment problem based on PDM. The experimental results show that the model and the indirect hybrid genetic algorithm has great effectiveness.更多还原