北京邮电大学学报

  • EI核心期刊

北京邮电大学学报 ›› 2015, Vol. 38 ›› Issue (1): 92-96.doi: 10.13190/j.jbupt.2015.01.018

• 研究报告 • 上一篇    下一篇

解决ACP高维优化问题的自适应多粒子模拟退火算法

啜钢1, 沈涛1, 罗海文2   

  1. 1. 北京邮电大学 信息与通信工程学院, 北京 100876;
    2. 中国联合网络通信有限公司 北京市分公司, 北京 100140
  • 收稿日期:2014-05-16 出版日期:2015-02-28 发布日期:2015-03-30
  • 作者简介:啜钢(1959—),男,教授,博士生导师,E-mail:Chuai@bupt.edu.cn.
  • 基金资助:

    国家科技重大专项项目(2013ZX03001003)

Adaptive Multi-Particles Simulated Annealing for High-Dimensional Optimization of ACP

CHUAI Gang1, SHEN Tao1, LUO Hai-wen2   

  1. 1. School of Information and Communication Engineering, Beijing University of Posts and Telecommunications, Beijing 100876, China;
    2. China United Network Communications Corporation Beijing Branch, Beijing 100140, China
  • Received:2014-05-16 Online:2015-02-28 Published:2015-03-30

摘要:

提出了一种改进的针对高维优化问题的自适应多粒子模拟退火(AMSA)算法,通过多个粒子对整个高维空间进行随机分割和相对独立的局部退火. 当每个局部于当前温度下达到稳态后,随着温度的降低,粒子依据自身状态和相互之间的关系自适应地减少粒子数目,以降低复杂度. 该算法用于解决通用移动通信系统自动小区规划问题. 仿真结果显示,对比其他用于解决高维优化问题的启发式算法,AMSA算法能在预定的时间内取得更理想的结果.

关键词: 模拟退火, 高维优化, 自适应多粒子, 库, 小区自动规划

Abstract:

An improved adaptive multi-particles simulated annealing (AMSA) was proposed to solve high-dimensional optimization problem. Thinking of multi particles, this new algorithm divides the entire high-dimensional space into several parts randomly as well as operates local annealing independently. Currently, when each local reaches steady state, with temperature decreasing, the particles are self-adaptively reduced for less complexity based on the relationship of the particles and their status. Consequently, AMSA can be used to settle high-dimensionally optimizing (universal mobile telecommunications system) automatic cell planning issue. Simulation shows that compared to other heuristic algorithms, it is good to settle high-dimensionally optimizing issue, and AMSA algorithm can achieve better results within a predetermined time.

Key words: simulated annealing, high-dimensional optimization, adaptive multi-particles, archive, automatic cell planning

中图分类号: