arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2005.09710math.NTmath.OCstat.CO
收藏

通过优化求解三个立方数之和

Sum of Three Cubes via Optimisation

Boian Lazov, Tsvetan Vetsov

AI总结:

本文将三个立方数之和问题转化为优化问题,应用粒子群优化和模拟退火等三种随机优化算法在 $k=2$ 情况下搜索整数解,并通过指数和对数正态分布建模运行时间,利用Fisher信息度量评估模型的可区分性。

摘要:

通过首先求解方程 $x^3+y^3+z^3=k$(固定 $k$)中的 $z$,然后考虑结果到最近整数函数的距离,我们将三个立方数之和问题转化为一个优化问题。然后,我们在 $k=2$ 的情况下将三种随机优化算法应用于此函数,该情况下有许多已知解。目标是测试该方法在搜索整数解时的有效性。这些算法是粒子群优化的一种修改版本和模拟退火的两种实现。我们希望比较它们的有效性,以算法的运行时间作为衡量标准。为此,我们通过假设两种底层概率分布——指数分布和对数正态分布来对时间数据建模,并计算它们的一些数值特征。最后,我们评估模型在具有相应 Fisher 信息度量的流形中关于测地距离的统计可区分性。

英文摘要:

By first solving the equation $x^3+y^3+z^3=k$ with fixed $k$ for $z$ and then considering the distance to the nearest integer function of the result, we turn the sum of three cubes problem into an optimisation one. We then apply three stochastic optimisation algorithms to this function in the case with $k=2$, where there are many known solutions. The goal is to test the effectiveness of the method in searching for integer solutions. The algorithms are a modification of particle swarm optimisation and two implementations of simulated annealing. We want to compare their effectiveness as measured by the running times of the algorithms. To this end, we model the time data by assuming two underlying probability distributions -- exponential and log-normal, and calculate some numerical characteristics for them. Finally, we evaluate the statistical distinguishability of our models with respect to the geodesic distance in the manifold with the corresponding Fisher information metric.

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。

↑