高级检索

    冗余及监控混合策略的优化配置算法研究

    Optimal Resources Allocation Algorithm for Optional Redundancy and Monitoring Strategies

    • 摘要: 大数据环境中监控和冗余混合策略的采用引起资源优化配置模型的状态空间膨胀,进化搜索算法在整型与非整型变量结合的解空间中的搜索效率有待提高,为此提出了基于搜索邻域分析的三元组模因算法.在分析了监控频率等参数变化对组件及系统可靠性增长影响的基础上,针对监控频率提出了基于变长邻域的近邻生成方法,针对策略选项提出了与组件关联的近邻生成方法.采用模因算法框架并改进了局部搜索算子,通过组件间的迭代搜索在保持个体优势的同时增大搜索范围.该算法能够用于求解混合策略下的组件保障措施选项及相应优化配置参数;与现有多策略搜索算法相比,在相同可靠性约束下,该算法能够得到消耗更低的资源配置结果;局部搜索策略对算法稳定性未造成明显影响.

       

      Abstract: In big data environment, the use of optional redundancy and monitoring strategy in one system increases the usage of resource and causes state space expansion for optimal resources allocation model. The performance of existing evolutionary search algorithms should be improved for the solution space formed by both integer and non-integer variables. To improve the algorithm efficiency, a memetic algorithm based on triple element array is proposed on the analysis of search neighborhood. First of all, the impact of change of variables such as monitoring rate on the system reliability increase is analyzed and then changing-length neighbor generation method is proposed for monitoring rate on neighbor analysis. The neighbor generation method is also proposed for strategy options considering the relations between components. After that, local search operator is refined through the iterative search among components, which increases the search range while maintaining the local advantage of individuals. This operator is used for improving the whole framework of memetic algorithm. Experiment results indicates that this algorithm can be used to get the solution of strategy option of each component and the corresponding optimized parameters for multiple optional strategies. Compared with existing multi-strategy search algorithms, the improved memetic algorithm could get better resources allocation results under the same reliability constraint. The local search operator does not have great impact on the stability of the whole algorithm.

       

    /

    返回文章
    返回