A function is called DC if it is expressible as the difference of two convex functions. In this work, we present a short tutorial on difference-of-convex optimization surveying and highlighting some important facts about DC functions, optimality conditions, and recent algorithms. The manuscript, accessible to a wide range of readers familiar with the convex analysis machinery, builds upon three pillars from variational analysis: directional derivative, ε-subdifferential, and the Legendre-Fenchel transform.

中文翻译：

---

DC编程基础知识

如果一个函数可以表示为两个凸函数之差，则称为DC。在这项工作中，我们提供了一个关于凸差优化测量的简短教程，并重点介绍了有关DC函数，最优性条件和最新算法的一些重要事实。该手稿是基于变量分析的三大支柱：方向导数，ε-次微分和Legendre-Fenchel变换，可供熟悉凸面分析机制的广泛读者使用。