Journal of Computer Applications ›› 2012, Vol. 32 ›› Issue (04): 1003-1008.DOI: 10.3724/SP.J.1087.2012.01003

• Advanced computing • Previous Articles     Next Articles

Spatial data fusion algorithm of CO2 based on cloud computing platform

HU Jun-guo,QI Heng-nian   

  1. College of Information Engineering, Zhejiang Agriculture and Forestry University, Lin’an Zhejiang 311300, China
  • Received:2011-10-31 Revised:2011-12-15 Online:2012-04-20 Published:2012-04-01
  • Contact: HU Jun-guo



  1. 浙江农林大学 信息工程学院,浙江 临安 311300
  • 通讯作者: 胡军国
  • 作者简介:胡军国(1978-),男,浙江临安人,副教授,硕士,CCF会员,主要研究方向:智能计算、分布式计算;祁亨年(1975-),男,浙江临安人,教授,博士,主要研究方向:模式识别、图像处理。
  • 基金资助:

Abstract: In order to fuse the massive CO2 dada, which were collected by the mobile sensor network from uncertain time and space, the paper analyzed the collected data. First, the test area was divided into m * n grids, ,and CO2 concentration was analyzed from each valid data of every grid. Second, according to the strong computing power of cloud computing, the paper put forward combined cloud model and design common clouds, breeding clouds, visual clouds and adjacent clouds. They ran relatively independently in the cloud and interacted with each cloud, forming distributed parallel computing system. Third, the paper modified the ants family, and designed common ants, breeding ants, visual ants and adjacent ants. All kinds of ants, which walked by their own rules, were assigned to different clouds and worked together harmoniously, with the pheromones and the optimal solution exchanging in local cloud and between global clouds by the cloud server. Finally, in Lingan of Zhejiang province the authors sampled 11080 data, and used Clounding V2 simulation platform to do a lot of experiments. The result shows that after searching 105 times the algorithm reaches stabilization, the optimization capability is 60 times as strong as the single algorithm, and that the ants in common clouds, breeding clouds, visual clouds and adjacent clouds are set 2∶2∶1∶1 can get the best performance.

Key words: cloud computing, ant colony algorithm, Ant Colony Optimization (ACO), CO2 analyzing, data fusion

摘要: 为了对移动传感器网络采集到的时间、空间不确定的海量CO2浓度数据进行融合,首先对采集的CO2数据进行分析,把测试区域分成m×n个网格,分析从每个网格取一个有效值来表示CO2浓度分布。然后根据云计算强大的计算能力,提出组合云模型,设计普通云、繁殖云、视觉云和邻接云,以云内相对独立运行和云间相互作用形成分布式并行计算机制。接着改造蚁群家族,设计普通蚂蚁、繁殖蚂蚁、视觉蚂蚁和邻接蚂蚁。各类蚂蚁分配到不同的云朵中,并按自身的规则运行,各类蚂蚁彼此配合工作,实现信息素和最优解在云内部局部交换和通过云服务器在云朵之间全局交换相结合。最后模拟生成有关临安的11080个数据,利用Clounding V2模拟平台进行大量实验,实验表明算法在105次寻优后基本趋于稳定,寻优能力是单机算法的60倍左右,并且普通云、繁殖云、视觉云和邻接云中的蚂蚁数量比设为2∶2∶1∶1性能表现出最佳。

关键词: 云计算, 蚁群优化, CO2分析, 数据融合, 并行计算

CLC Number: