Abstract: As to the reasonable matching between the resources needed by multi-task and M-dimensional resources offered by multiprocessing nodes, the author offered multiple dimension set partitioning optimization models, defined the function of resource balancing degree and proposed a Multiple Dimension Set Partitioning Load Balancing Resource Optimization Allocation Algorithm(MDSPLBROAA), which can optimize to solve the NP problems. The experimental results presented show that the algorithm has better practicability and feasibility. Furthermore, it has higher efficiency than the traditional heuristic algorithm.

Key words: load balancing, resource optimization, resource balancing degree, multiple dimension set partitioning

摘要： 针对多处理群集系统中多个任务处理需要的资源和多个处理节点能够提供的m维资源间的合理匹配问题，给出了多维集合划分问题的优化模型，定义了资源均衡度函数，提出多维集合划分负载均衡资源优化分配算法，通过该算法可以得到资源匹配NP问题的较优解。实验结果表明，该算法具有较好的实用性和可行性，比传统的启发式算法效率高。

关键词: 负载均衡, 资源优化, 资源优化, 多维集合划分