北京邮电大学学报

  • EI核心期刊

北京邮电大学学报 ›› 2019, Vol. 42 ›› Issue (1): 109-113.doi: 10.13190/j.jbupt.2018-036

• 研究报告 • 上一篇    下一篇

一种适用于(p+m)-中点问题的服务设施放置算法

焦计平, 逯海, 洪学敏, 石江宏   

  1. 厦门大学 信息科学与技术学院, 厦门 261005
  • 收稿日期:2018-02-07 出版日期:2019-02-28 发布日期:2019-03-08
  • 作者简介:焦计平(1985-),男,博士生,E-mail:jjp@xmu.edu.cn;石江宏(1968-),男,教授,博士生导师.
  • 基金资助:
    国家自然科学基金项目(61571378)

A Service Facility Placement Algorithm for the (p+m)-Median Problem

JIAO Ji-ping, LU Hai, HONG Xue-min, SHI Jiang-hong   

  1. School of Information Science and Engineering, Xiamen University, Xiamen 361005, China
  • Received:2018-02-07 Online:2019-02-28 Published:2019-03-08
  • Supported by:
     

摘要: 针对雾计算应用中服务设施放置问题,将其建模成(p+m)-中点问题,提出了一种基于贪婪策略与禁忌搜索策略相结合的启发式服务设施放置算法.提出的算法适用于一般拓扑、任意需求分布的网络.性能分析结果表明,提出的算法是多项式时间的,在当扩展服务节点数和请求节点数相等时能够达到性能上的最优.仿真结果验证了新算法的有效性.

关键词: 放置问题, 禁忌搜索, 搜索域, 图状网络

Abstract: This service facility placement problem is investigated, which appears representatively in fog computing for the cost minimization and the optimal resource utilization. After the problem being modelled as a (p+m)-median problem, a novel heuristic placement algorithm is proposed that combines the greedy and tabu-search strategies. The proposed algorithm can be used in networks with arbitrary topology and random demand distribution. Analysis results show that it is polynomial in time complexity and can reach the optimal performance in the case that the number of the extended service nodes in the network is equal to that of the request nodes. Finally, simulations verify the advantages above.

Key words: placement problem, tabu-search, search domain, graph network

中图分类号: