CAJ | 학술논문

由于云计算的动态性、异构性和不可预测性等特点，使得资源调度策略面临很大的挑战。目前解决资源调度的方法主要是一些启发式算法，如模拟退火算法、人工神经网络算法、粒子群算法、蚁群算法和遗传算法等，由于优缺点分明，不能单独实现云计算任务的最优分配。因此，提出了使用混合优化算法解决云计算资源分配问题。在算法前期，借助粒子群全局广泛搜索能力，快速寻找到较优解；在算法后期，借助蚁群算法的正反馈性和高效性，寻找最优解。实验表明该算法有较短的任务执行时间和实现各个物理主机间的负载均衡。
유우운계산적동태성、이구성화불가예측성등특점，사득자원조도책략면림흔대적도전。목전해결자원조도적방법주요시일사계발식산법，여모의퇴화산법、인공신경망락산법、입자군산법、의군산법화유전산법등，유우우결점분명，불능단독실현운계산임무적최우분배。인차，제출료사용혼합우화산법해결운계산자원분배문제。재산법전기，차조입자군전국엄범수색능력，쾌속심조도교우해；재산법후기，차조의군산법적정반궤성화고효성，심조최우해。실험표명해산법유교단적임무집행시간화실현각개물리주궤간적부재균형。
It makes resource scheduling policy a big challenge because of the dynamic nature of cloud computing, heterogeneous and unpredictable characteristics. The present solution are heuristic algorithms, such as simulated annealing, artificial neural network algorithm, particle swarm optimization, ant colony optimization, genetic algorithm and so on;It cannot be achieved optimal allocation of cloud computing tasks separately due to all these methods have its advantages and disadvantages. So this study try to fix the problem of resource scheduling of cloud computing using Hybrid optimization algorithm. In the early stage of algorithm, using a wide range global search capability of Particle Swarm Optimization to find the optimum solution quickly;In the late stage, with positive and efficiency of feedback Ant Colony Algorithm, the optimal solution is found. Experimental results show that task execution time of the algorithm is shorter and make load balancing for each physical host.