摘要
云计算环境下应用蚁群算法分布式并行对问题进行求解的研究较少,且蚁群算法存在搜索时间长和易收敛于非最优解的缺陷,当问题的规模较大时求解困难。为此应用云计算技术将蚁群算法并行化,提出基于MapReduce的蚁群算法。该算法将分治思想和模拟退火算法融入蚁群算法,改进其缺陷,并应用于求解较大规模的旅行商问题。仿真实验取得了较好的效果,且获得了测试实例gr666的新解。
The researches on solving problems with Ant Colony Optimization(ACO)distributed parallel under cloud computing were less, and ACO had defects in long search time and convergence in non-optimal solution. When the scale of problem was large, it was too hard to solve. Therefore, MapReduce-based ACO was proposed by using cloud computing to parallel ACO. In this algorithm, dividing conquer and simulated annealing algorithm were merged into ACO to improve the defects. It was also applied to solve large-scale of Traveling Salesman Problem(TSP). The simulation experiment got a well effect and the new solutions of test gr666 were obtained.
出处
《计算机集成制造系统》
EI
CSCD
北大核心
2012年第7期1503-1509,共7页
Computer Integrated Manufacturing Systems
基金
国家863计划资助项目(2011AA040501)
国家社会科学基金资助项目(10CGL024)
安徽省教育厅自然科学重点资助项目(KJ2011A006)~~
关键词
云制造
云计算
蚁群算法
分治
模拟退火算法
旅行商问题
cloud manufacturing
cloud computing
ant colony optimization
divide conquer
simulated annealing algorithm
traveling salesman problem