关于线性规划问题熵障碍对偶法的注记

线性规划是目标优化问题中最常用的模型。关于大规模线性规划问题的有效求解问题一直受到人们的关注。熵障碍对偶法是继内点法之后,又一解线性规划问题的新的算法。本文讨论了熵障碍对偶法的推广形式及其梯度类算法的收敛性。

A Note on Dual Entropic Barrier of Linear Programming

In goal-seeking optimization problems, linear programming problems are the most common. The efficient algorithms for large-scale linear programming problem have been attracting considerable attention from both theorists and practitioners in a variety of disciplines since 1970s. After interior-point methods, the dual entropic barrier method is another new algorithm. This paper gives the extension of dual entropic barrier method and the convergence of its gradient method is discussed.