The standard measure for the quality of online algorithms is the competitive ratio. This measure is generally applicable, and for some problems it works well, but for others it fails to distinguish between algorithms that have very different performance. Thus, ever since its introduction, researchers have worked on improving the measure, defining variants, or defining measures based on other concepts to improve on the situation. Relative worst-order analysis (RWOA) is one of the most thoroughly tested such proposals. With RWOA, many separations of algorithms not obtainable with competitive analysis have been found. In RWOA, two algorithms are compared directly, rather than indirectly as is done in competitive analysis, where both algorithms are compared separately to an optimal offline algorithm. If, up to permutations of the request sequences, one algorithm is always at least as good and sometimes better than another, then the first algorithm is deemed the better algorithm by RWOA. We survey the most important results obtained with this technique and compare it with other quality measures. The survey includes a quite complete set of references.

中文翻译：

---

相对最差分析

在线算法质量的标准衡量标准是竞争比率。这个度量是普遍适用的，对于某些问题它工作得很好，但对于其他问题，它无法区分性能差异很大的算法。因此，自推出以来，研究人员一直致力于改进测量、定义变体或基于其他概念定义测量以改善情况。相对最坏阶分析 (RWOA) 是经过最彻底测试的此类建议之一。使用 RWOA，已经发现了许多通过竞争分析无法获得的算法分离。在 RWOA 中，直接比较两种算法，而不是像在竞争分析中那样间接比较两种算法，其中两种算法分别与最佳离线算法进行比较。如果，直到请求序列的排列，一个算法总是至少和另一个算法一样好，有时甚至更好，那么第一个算法被 RWOA 认为是更好的算法。我们调查了使用这种技术获得的最重要的结果，并将其与其他质量测量方法进行比较。该调查包括一套相当完整的参考资料。