二次规划子问题的一种非精确光滑牛顿法  

An Inexact Smoothing Newton Method for Quadratic Programming Subproblem

作　　者：朱子旋 芮绍平[1] 蔡玉玉 ZHU Zixuan;RUI Shaoping;CAI Yuyu(School of Mathematical Sciences,Huaibei Normal University,235000,Huaibei,Anhui,China)

机构地区：[1]淮北师范大学数学科学学院,安徽淮北235000

出　　处：《淮北师范大学学报（自然科学版）》2023年第2期21-25,共5页Journal of Huaibei Normal University：Natural Sciences

基　　金：安徽省高等学校自然科学研究项目(KJ2020A0024);淮北师范大学实验室开放项目(2022sykf016)。

摘　　要：二次规划子问题的求解是解决规划问题的关键。针对二次规划子问题,利用最优性条件,借助光滑逼近函数将其转化为光滑方程组,结合非精确牛顿法得到一种求解二次规划子问题的非精确光滑牛顿法。一定条件下证明其全局收敛性。数值实验表明此算法对二次规划子问题有效。The solution of the quadratic programming sub problem is the key to solving the planning problem.In this paper,an inexact smooth Newton method for solving the quadratic programming sub problem is ob⁃tained by transforming it into a smooth system of equations with the help of smooth approximation functions using optimality conditions and combining it with an inexact Newton method.Its global convergence is proved under certain conditions.Numerical experiments show that this algorithm is effective for quadratic program⁃ming sub problems.

关 键 词：二次规划子问题 全局收敛性 非精确牛顿法 

分 类 号：O221.2[理学—运筹学与控制论]