合理的资源调度可以在很大程度上提高网格系统资源的利用率,遗传算法(GA)因具有强大稳健的隐并行解空间搜索功能,被广泛应用于任务分配和调度问题的求解。本文在研究标准遗传算法(SGA)的基础上,提出与小生境技术相结合的自适应选择概率、父子竞争(PCC)交叉算子、插入变异算子和最优保存策略,改进SGA算法,在很好地保持种群收敛性的同时,提高了算法的局部和全局搜索能力。仿真实验结果表明,本文算法与其它调度算法比较,更能有效地实现资源的分配,可以成功应用于网格环境下独立任务的分配与调度。
Reasonable resource scheduling can greatly improve the utilization of the grid.Genetic algorithm(GA) for powerful and implicit parallel space search capability is widely used to solve task allocation and scheduling problems.Based on the research on existing scheduling algorithms,this paper describes the adaptive selection probability combined with niche technology,PCC(parents and children competition) crossover operator,insert mutation operator and elitist strategy to improve GA,it keeps the population's convergence and increases the efficiency of local and global search capability.Simulation results show that this algorithm is more effective for the allocation of resources compared with other algorithm,it can be successfully applied to independent task allocation and scheduling in grid.