Abstract: This paper uses Level-Set-Based Particle Swarm Optimizion(LSBPSO) to solve the problem of disk load balance. The characteristic of convergence of the algorithm swiftly adjusts the mapping between logic disk and physic disk which uses the technique of disk striping. It refers a new method to forecast the hot of logic disk and adds the content of expression of hot of physic disk. Simulation shows that the algorithm is efficient and defective.

Key words: Particle Swarm Optimization(PSO), disk load, technique of disk striping

摘要： 将基于水平集的粒子优化算法(LSBPSO)应用于磁盘负载均衡问题中，利用LSBPSO的快速收敛性动态调整分条技术下逻辑磁盘和物理磁盘的映射关系。提出一种新的逻辑磁盘热度预测方法，对物理磁盘的热度的表示方法进行扩充。实验表明，LBPSO能较好地解决分条技术下的磁盘负载问题，收敛速度较快。

关键词: 粒子群优化算法, 磁盘负载, 分条技术

CLC Number: 

• TP301.6