基于多分辨率聚类的云制造任务分配  

Task allocation in cloud manufacturing based on multi-resolution clustering

作　　者：戈鹏[1] 杨欣[1] 肖雄辉[1] 邱厌庆[1] 任佩瑜[1] 刘柱胜[1] 郑伟民[1] 

机构地区：[1]四川大学工商管理学院,四川成都610064

出　　处：《计算机集成制造系统》2012年第7期1461-1468,共8页Computer Integrated Manufacturing Systems

基　　金：国家863计划重大资助项目(2008AA04A107);国家自然科学基金重大国际合作资助项目(71020107027);四川大学985工程资助项目;九寨沟风景名胜区管理局博士后科研工作站资助项目~~

摘　　要：针对云制造环境下用户需求复杂多变和制造资源能力分散异构的特点,为实现制造资源的动态共享与智能分配,提出一种能够在云制造环境下迅速得出需求能力高效匹配的多分辨率聚类任务分配方案。根据任务目标特点选择适应的参数建立反映需求能力匹配程度的聚类距离函数,并以此距离为聚类依据展开多分辨率聚类,随聚类过程将任务分解并分配到相应的资源能力提供者,得出云制造系统中需求能力的多分辨率层次结构,为解决云制造环境下资源的高效配置问题、实现云制造系统中用户需求和资源能力的动态匹配提供了可行结构。通过实例对以上任务分配方案过程进行了详述,并论证了该算法可为充分利用云制造系统中零碎制造能力提供有效途径。Aiming at the characteristics of users＇ complicated variable requirements and manufacturing resources＇ dispersed heterogeneous in cloud manufacturing environment, a multi-resolution clustering task allocation method which could get efficient match of demand ability, was proposed to achieve dynamic sharing and intelligent distribution of manufacturing resources. According to characteristic of task goal, appropriate parameters were selected to build function of clustering distance, and the multi-resolution clustering was developed based on this distance. Tasks were decomposed and distributed to corresponding resources along with the clustering process, and a hierarchical structure of demand ability in cloud manufacturing system was obtained. It was a feasible structure to solve the problem of resource efficient allocation and to realize the dynamic matching between requirements and resource ability in cloud manufacturing system. Through examples, the process of proposed allocation scheme was illustrated, and it was proved to be an effective way to make full use of broken manufacturing capacity in cloud manufacturing.

关 键 词：云制造 任务分配 多分辨率聚类 资源匹配程度 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TH166[自动化与计算机技术—计算机科学与技术]