We present an $O(\log^3\log n)$-round distributed algorithm for the $(\Delta+1)$-coloring problem, where each node broadcasts only one $O(\log n)$-bit message per round to its neighbors. Previously, the best such broadcast-based algorithm required $O(\log n)$ rounds. If $\Delta \in \Omega(\log^{3} n)$, our algorithm runs in $O(\log^* n)$ rounds. Our algorithm's round complexity matches state-of-the-art in the much more powerful CONGEST model [Halld\'orsson et al., STOC'21 & PODC'22], where each node sends one different message to each of its neighbors, thus sending up to $\Theta(n\log n)$ bits per round. This is the best complexity known, even if message sizes are unbounded. Our algorithm is simple enough to be implemented in even weaker models: we can achieve the same $O(\log^3\log n)$ round complexity if each node reads its received messages in a streaming fashion, using only $O(\log^3 n)$-bit memory. Therefore, we hope that our algorithm opens the road for adopting the recent exciting progress on sublogarithmic-time distributed $(\Delta+1)$-coloring algorithms in a wider range of (theoretical or practical) settings.
翻译:我们提出了一个 $O(\log^3\log n)$ 轮的分布式算法,用于 $(\Delta+1)$-着色问题,其中每个节点每轮仅向其邻居广播一个 $O(\log n)$ 位的消息。此前,最好的基于广播的算法需要 $O(\log n)$ 轮。如果 $\Delta \in \Omega(\log^{3} n)$,我们的算法可以在 $O(\log^* n)$ 轮内运行。该算法的轮复杂度与在更强大的 CONGEST 模型 [Halld\'orsson et al., STOC'21 & PODC'22] 中的最新状态相匹配,其中每个节点向每个邻居发送不同的消息,因此每轮发送最多 $\Theta(n\log n)$ 位。即使消息大小不受限制,这仍是已知的最优复杂度。我们的算法足够简单,可以在更弱的模型中实现:如果每个节点以流式读取方式读取其接收到的消息,并使用仅 $O(\log^3 n)$ 位的存储器,则可以实现相同的 $O(\log^3\log n)$ 轮复杂度。因此,我们希望我们的算法能够以更广泛的(理论或实际)设置中采用子对数时间分布式 $(\Delta+1)$-着色算法的最新进展。