基于模糊聚类和遗传算法的制造资源优化配置方法研究

发布时间：2018-05-13 03:31

本文选题：云制造 + 制造资源　；参考：《西安理工大学》2017年硕士论文

【摘要】：近年来,云制造等先进网络化制造模式的出现,推动着制造业向全球化、一体化的方向发展。在云模式下,制造资源的优化配置问题是实现“制造即服务”理念的关键。本文研究制造资源的服务化封装、聚类分析以及优化配置等,对于提升云制造的服务效率和大规模普及具有重要的现实意义。建立制造资源服务化描述模型与虚拟化封装机制。针对制造资源海量性、异构性及独立性的特征,基于面向对象的思维构建制造资源服务化描述模型,对模型中各模块的功能和意义进行讨论。针对资源结构和语义等验证问题,建立一种基于Web的资源虚拟化封装框架。对制造资源进行模糊聚类研究。针对制造资源间关联性差、不利于实现高效检索和优化配置的问题,分别运用基于模糊相似矩阵的直接聚类法和FCM聚类算法对某一批制造资源进行聚类分析,对算法中的属性权值设定、初始聚类中心的选取等问题进行讨论和改进,提出了一种将两种聚类方法进行结合运用的新算法。改进制造资源优化配置算法。针对云模式下制造资源的优化配置问题,建立相关问题的评价函数模型,对传统遗传算法进行自适应改进,采用混合轮盘赌法和锦标赛法的选取机制提高了算法的寻优效率;采用自适应交叉率和变异率的方法保证了算法的多样性和收敛性,赋予变异过程正反馈特性,提高了算法的稳定性和求解质量。最后以实例验证了算法改进的有效性,并得出了最优的资源配置方案。设计并实现了制造资源优化配置原型系统。根据制造资源配置研究中的诸多需求,以云制造服务平台为研究主体,对平台的总体结构、各模块功能、理论和技术等进行设计与开发,验证了本文所研究理论与方法的可行性与有效性。
[Abstract]:In recent years, the emergence of advanced networked manufacturing model such as cloud manufacturing promotes the development of globalization and integration of manufacturing industry. In the cloud mode, the optimal allocation of manufacturing resources is the key to realize the concept of "manufacturing as a service". In this paper, the service encapsulation of manufacturing resources, clustering analysis and optimization of configuration are studied, which is of great practical significance for improving the service efficiency of cloud manufacturing and popularizing it on a large scale. Establish manufacturing resource service description model and virtualization encapsulation mechanism. Aiming at the characteristics of magnanimity, heterogeneity and independence of manufacturing resources, a service-oriented description model of manufacturing resources is constructed based on object-oriented thinking, and the functions and significance of each module in the model are discussed. A resource virtualization encapsulation framework based on Web is proposed for resource structure and semantics verification. Research on fuzzy clustering of manufacturing resources. Aiming at the problem of poor correlation among manufacturing resources, which is not conducive to efficient retrieval and optimal configuration, the direct clustering method based on fuzzy similarity matrix and FCM clustering algorithm are applied to cluster analysis of a batch of manufacturing resources. This paper discusses and improves the setting of attribute weight and the selection of initial clustering center in the algorithm, and proposes a new algorithm which combines the two clustering methods. Improve the optimal allocation algorithm of manufacturing resources. Aiming at the optimal allocation of manufacturing resources in cloud mode, the evaluation function model of related problems is established, and the traditional genetic algorithm is improved adaptively. The selection mechanism of mixed roulette method and tournament method is adopted to improve the efficiency of the algorithm. The adaptive crossover rate and mutation rate are used to ensure the diversity and convergence of the algorithm, and the positive feedback characteristics of the mutation process are given to improve the stability and quality of the algorithm. Finally, an example is given to verify the effectiveness of the improved algorithm, and the optimal resource allocation scheme is obtained. A prototype system for optimal configuration of manufacturing resources is designed and implemented. According to many requirements in the research of manufacturing resource allocation, this paper designs and develops the overall structure, function of each module, theory and technology of the platform, taking the cloud manufacturing service platform as the main body. The feasibility and validity of the theory and method studied in this paper are verified.
【学位授予单位】：西安理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP18;TP393.09

【参考文献】