arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1803.04744cs.DS
收藏

论可编程、离散化和卷积

On Integer Programming, Discrepancy, and Convolution

Klaus Jansen, Lars Rohwedder

摘要:

当m为固定常数时,具有m个约束的可解规划在伪多项式时间内可解于约束中的最大系数$Δ$。我们给出了一个新的算法的运行时间为$O(\sqrt{m}Δ)^{2m} + O(nm)$,这是最先进的改进。此外,我们证明了我们的算法的任何$m$的改进是等效于改进的二次时间算法的$(\min,~+)$卷积。这是一个强有力的证据,表明我们的算法的运行时间是最好的。我们还提出了一个专门的算法与运行时间$O(\sqrt{m} Δ)^{(1 + o(1))m} + O(nm)$测试的整数规划的可行性,也给出了一个紧的下限,这是基于SETH在这种情况下。

英文摘要:

Integer programs with m constraints are solvable in pseudo-polynomial time in $Δ$, the largest coefficient in a constraint, when m is a fixed constant. We give a new algorithm with a running time of $O(\sqrt{m}Δ)^{2m} + O(nm)$, which improves on the state-of-the-art. Moreover, we show that improving on our algorithm for any $m$ is equivalent to improving over the quadratic time algorithm for $(\min,~+)$-convolution. This is a strong evidence that our algorithm's running time is the best possible. We also present a specialized algorithm with running time $O(\sqrt{m} Δ)^{(1 + o(1))m} + O(nm)$ for testing feasibility of an integer program and also give a tight lower bound, which is based on the SETH in this case.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑