We study the arbitrary cost case of the unweighted Stochastic Score Classification (SSClass) problem. We show two constant approximation algorithms and both algorithms are 6-approximation non-adaptive algorithms with respect to the optimal adaptive algorithm. The first algorithm uses a modified round-robin approach among three sequences, which is inspired by a recent result on the unit cost case of the SSClass problem. The second algorithm is originally from the work of Gkenosis et al. In our work, we successfully improve its approximation factor from 2(B-1) to 6. Our analysis heavily uses the relation between computation and verification of functions, which was studied in the information theory literature.
翻译:我们研究了未经加权的斯托卡分级(SSClass)问题的任意成本案例。我们显示了两种恒定近似算法,这两种算法在最佳适应算法方面均为6-接近非适应性算法。第一种算法在三个序列中采用经修改的圆环路对数法,这受到SSClas问题单位成本案例最近结果的启发。第二种算法最初来自Gkenosis等人的工作。在我们的工作中,我们成功地将其近似系数从2(B-1)提高到了6。我们的分析大量利用了计算和核实功能之间的关系,信息理论文献对此进行了研究。