Title: Minimizing Sum of Truncated Convex Functions and its Applications

Abstract: We study a class of problems where the sum of truncated convex functions is minimized. In statistics it is often encountered when $L_0$-penalized models are fitted. While in general it leads to an NP-Hard non-convex optimization problem, we propose a general strategy for the exact solution with finite computing time. In particular, with functions whose minima possess closed form solution such as quadratic functions, there are polynomial-time algorithms in low-dimensional space. In simulation studies with the sum the truncated quadratic functions, our algorithm has superior performance when compared with other global optimization algorithms, especially when the objective function assumes a complex landscape. We also demonstrate the utility of our algorithm for outlier detection in simple linear regression, where it shows strength under the presence of many outliers.