摘要
针对约束非线性l_1问题不可微的特点,提出了一种光滑近似算法.该方法利用"+"函数的光滑近似函数和罚函数技术将非线性l_1问题转化为无约束可微问题,并在适当的假设下,该算法是全局收敛的.初步的数值试验表明算法的有效性.
A smooth approximating algorithm is proposed to study the non-differenti- able constrained nonlinear 11 problem.Based on the smooth approximating function of the plus function and penalty function,the constrained nonlinear l_1 problem is converted to an unconstrained differentiable problem.The algorithm is global convergent under mild conditions.Primary numerical results demonstrate that the algorithm is efficient.
出处
《应用数学与计算数学学报》
2007年第2期97-101,共5页
Communication on Applied Mathematics and Computation
基金
南通大学自然科学基金(07Z002)
关键词
非线性ι1问题
光滑近似函数
收敛性
nolinear l_1 problem
smooth approximating function
convergence
