OUTLIER DETECTION IN TIME SERIES VIA MIXED-INTEGER CONIC QUADRATIC OPTIMIZATION
OUTLIER DETECTION IN TIME SERIES VIA MIXED-INTEGER CONIC QUADRATIC OPTIMIZATION
复制标题
DOI：
10.1137/19m1306233
复制
发表时间：
2021-01-01
影响因子：
3.1
通讯作者：
Gomez, Andres
中科院分区：
文献类型：
--
作者：
Gomez, Andres
我们考虑了在给定被异常值破坏的噪声观测值的情况下估计维纳过程真值的问题。本文给出了如何改进现有的混合整数二次优化公式来解决这一问题。具体地说，我们通过提升对已有的公式进行了凸化，得到了新的混合整数二次元的重新表达式。当与当前的混合整数优化求解器一起使用时，所提出的重新公式更强大，并且实质上更快。在我们的实验中，解决时间至少提高了两个数量级。
We consider the problem of estimating the true values of a Wiener process given noisy observations corrupted by outliers. In this paper we show how to improve existing mixed-integer quadratic optimization formulations for this problem. Specifically, we convexify the existing formulations via lifting, deriving new mixed-integer conic quadratic reformulations. The proposed reformulations are stronger and substantially faster when used with current mixed-integer optimization solvers. In our experiments, solution times are improved by at least two orders-of-magnitude.