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 2110.09901math.OC
收藏

实数子集和问题的一个FPTAS(完全多项式时间近似方案)

A FPTAS for the Subset Sum Problem with Real Numbers

Marius Costandin

AI总结:

本文针对实数子集和问题,将其转化为单位超立方体与两半空间相交凸集上的固定点距离最大化问题,提出多项式算法,可在任意精度下通过求解最大化点判定问题可行性,给出该问题的FPTAS。

摘要:

本文研究实数域上的子集和问题。我们从给定问题出发,构建了一个定义在多胞体上的二次最大化问题,该问题最终可转化为到某一固定点的距离最大化问题。为求解该问题,我们提出了一种多项式算法,用于在特定凸集上最大化到固定点的距离;该凸集由单位超立方体与两个相关半空间相交得到。我们证明,若子集和问题存在解,我们的算法能以任意选定的精度给出正确的最大距离。在此情形下,求得的最大化点就是子集和问题的一个解。因此,我们通过计算该最大化点并对其分析,即可判定子集和问题的可行性。

英文摘要:

In this paper we study the subset sum problem with real numbers. Starting from the given problem, we formulate a quadratic maximization problem over a polytope which is eventually written as a distance maximization to a fixed point. For solving this, we provide a polynomial algorithm which maximizes the distance to a fixed point over a certain convex set. This convex set is obtained by intersecting the unit hypercube with two relevant half spaces. We show that in case the subset sum problem has a solution, our algorithm gives the correct maximum distance up to an arbitrary chosen precision. In such a case, we show that the obtained maximizer is a solution to the subset sum problem. Therefore, we compute the maximizer and upon analyzing it we can assert the feasibility of the subset sum problem.

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。

↑