In recent years, there has been a growing interest in solving various graph coloring problems in the streaming model. The initial algorithms in this line of work are all crucially randomized, raising natural questions about how important a role randomization plays in streaming graph coloring. A couple of very recent works have made progress on this question: they prove that deterministic or even adversarially robust coloring algorithms (that work on streams whose updates may depend on the algorithm's past outputs) are considerably weaker than standard randomized ones. However, there is still a significant gap between the upper and lower bounds for the number of colors needed (as a function of the maximum degree $\Delta$) for robust coloring and multipass deterministic coloring. We contribute to this line of work by proving the following results. In the deterministic semi-streaming (i.e., $O(n \cdot \text{polylog } n)$ space) regime, we present an algorithm that achieves a combinatorially optimal $(\Delta+1)$-coloring using $O(\log{\Delta} \log\log{\Delta})$ passes. This improves upon the prior $O(\Delta)$-coloring algorithm of Assadi, Chen, and Sun (STOC 2022) at the cost of only an $O(\log\log{\Delta})$ factor in the number of passes. In the adversarially robust semi-streaming regime, we design an $O(\Delta^{5/2})$-coloring algorithm that improves upon the previously best $O(\Delta^{3})$-coloring algorithm of Chakrabarti, Ghosh, and Stoeckl (ITCS 2022). Further, we obtain a smooth colors/space tradeoff that improves upon another algorithm of the said work: whereas their algorithm uses $O(\Delta^2)$ colors and $O(n\Delta^{1/2})$ space, ours, in particular, achieves (i)~$O(\Delta^2)$ colors in $O(n\Delta^{1/3})$ space, and (ii)~$O(\Delta^{7/4})$ colors in $O(n\Delta^{1/2})$ space.
翻译:近些年来,人们越来越有兴趣解决 流式模型中各种图表颜色的问题。 然而, 这行的初始算法都非常随机化, 引起了关于流式图形颜色颜色中随机化作用的重要性的自然问题。 一些最近的工作在这个问题上取得了进展: 它们证明, 确定性甚至敌对性强的颜色算法( 这些流的更新可能取决于算法过去的产出) 大大弱于标准随机化的数据。 然而, 在所需的颜色数量的上下限之间仍然存在着巨大的差距 : 最强的颜色设计 $\Delta$\D$。 我们通过证明以下结果来推动这项工作。 在确定性半流的半流中( $( n)\dot\ text{poly} n) 空间制度中, 我们的变色性算法( =D$+1美元) 的上下限值之间仍然存在巨大的差距 。