作 者: (曾琴);
机构地区: 重庆师范大学数学科学学院,重庆401331
出 处: 《重庆理工大学学报(自然科学)》 2017年第8期182-191,共10页
摘 要: 带有线性约束条件且目标函数是块可分的凸性最小问题一直是研究的重点。该问题是经过研究目标函数由2个不是充分光滑的凸函数组成或是由3个不是充分光滑的凸函数组成,从而推广到目标函数由n个不是充分光滑的凸函数组成的情况。问题的解决是以经典的交替方向法为基础,延伸出多种方法来建立该模型。总结了几种常见方法,同时提出了新方法——基于分离法的新ADMM法,并证明该方法的可行性。 The convexity minimum problem with a linear constraints and a block separable objective function is always studied. As you can see from this article,the objective function which is composed of two or three insufficiently smooth convex functions is firstly studied to promote the objective function composed of n insufficiently smooth convex functions. The solution to the model is based on the classic alternating direction method. Therefore,it stretches out a lot of kinds of methods to solve the model.The article summarizes the common methods,and then it puts forward a new method: a new ADMM method based on separation process,and it simultaneously indicates the method's feasibility.