A Method of Reduced. Subgradient for Nondifferentiable Convex Programming uuder Linear Constraints
线性约束不可微凸规划的既约次梯度法

Making use of part of the information about the subdifferential of the object function,this paper presents an algorithm of reduced subgradient for the nondifferentiable convex pro-gramming. The theoretical model for proving the convergence of the algorithm is also con-tructed. The global convergence of the algorithm is proved and feasibility of the algorithm isdemonstrated.