基于重建策略的云工作流调度算法优化

为了进一步提高算法性能,提出一种改进的蛙跳算法,并与调度方案相结合,以期为云工作流资源分配提供最优调度.通过在蛙跳算法的局部搜索中加入重建策略,提高了数据随机性,有效避免了局部最优.研究了调度方案生成算法,与改进算法相结合得到接近最优的调度.利用Java模拟器进行仿真试验,并与粒子群优化算法和传统蛙跳算法作比较.实验证明,提出的方法可以在满足最长截止时间约束的情况下,使总执行成本最小化.

Optimization of cloud workflow scheduling algorithm based on reconstruction strategy

In order to further improve the performance of the algorithm, this paper proposes an improved leapfrog algorithm, which is combined with the scheduling scheme to provide optimal scheduling for cloud workflow resource allocation. First of all, by means of joining the reconstruction strategy to improve the randomness of data the local search in the frog leaking algorithm the local optimal is effectively prevented. Secondly, we study the scheduling scheme generation algorithm, and get the optimal scheduling with the improved algorithm. Finally, the simulation experiment is carried out using Java simulator, and compared with the particle swarm optimization algorithm and the traditional frog leap algorithm. It is found that the proposed method can minimize the total execution cost while satisfying the longest deadline constraint.