网格负载均衡策略及其蚁群优化算法

Load balancing strategy and ant optimization algorithm for grids

以重庆大学CampusGrid建设和加入ChinaGrid的发展规划为背景,研究了多网格环境中出现共用节点(即同时为多个网格系统服务的节点)时资源利用率下降问题,并针对该问题提出了以提高资源利用率为优化目标的负载均衡算法。主要分为问题模型建立、算法设计、以及实验评估3个部分。提出的算法能较好解决该问题,并考虑了网络通信开销对算法执行效果的影响。实验表明,提出的算法能有效防止网格中出现共用节点时资源利用率的下降,并对网格动态变化的特性具有较强的适应能力。

In view of the Campus Grid construction,which is also a crucial part of ChinaGrid project,the performance decline for grid scheduling algorithms when non-dedicated nodes emerge in multi-grid environment is studied. A load balancing algorithm to optimize resource usage rate is proposed. The paper involves three parts：problem modeling,algorithm design,and experiment evaluation. The experimental results show that the proposed algorithm is effective for solving the problem of resource usage rate decline under the discussed grid circumstance.