CAJ | 학술논문

针对云计算中的虚拟机放置问题，选取负载均衡和能源消耗两个目标进行多目标优化。利用改进多目标优化蚁群算法求解最优解集。针对文章定义的负载均衡因子和能源消耗归一化函数，重新定义了蚁群算法的信息素、启发素、信息素更新规则、状态转移规则。利用 Cloudsim 平台对算法性能进行仿真，并与降序首次适应算法、分组遗传算法、针对能源消耗的单目标优化蚁群算法进行对比。结果表明该算法能更好地提升系统负载均衡性能和减少能源消耗，实现多目标优化。
침대운계산중적허의궤방치문제，선취부재균형화능원소모량개목표진행다목표우화。이용개진다목표우화의군산법구해최우해집。침대문장정의적부재균형인자화능원소모귀일화함수，중신정의료의군산법적신식소、계발소、신식소경신규칙、상태전이규칙。이용 Cloudsim 평태대산법성능진행방진，병여강서수차괄응산법、분조유전산법、침대능원소모적단목표우화의군산법진행대비。결과표명해산법능경호지제승계통부재균형성능화감소능원소모，실현다목표우화。
Aiming at virtual machine placement problem in the cloud computing environment ,load balancing and power consumption were selected as two objectives for multi-objective optimization .An extended Multi-Objective Ant Colony Opti-mization(MACO) algorithm was used to find the best solution set .According to the load balancing factor and power con-sumption normalized function defined in this paper ,the pheromone trail ,heuristic information ,pheromone trail update rules and state transition rules of MACO were redefined .The algorithm performance was simulated on Cloudsim and compared with FFD ,Grouping Genetic Algorithm and power consumption Single-Objective Ant Colony Optimization algorithm .The results showed that the MACO algorithm can effectively improve the system load balancing performance ,reduce the power consumption and realize the Multi-Objective Optimization .