# zbMATH — the first resource for mathematics

A reduced subgradient algorithm. (English) Zbl 0624.90085
The authors describe an iterative method for the approximate solution of nondifferentiable convex programming problems. The problems incorporate linear equality and inequality constraints $$(Ax=b$$, $$x\geq 0)$$; the method combines Wolfe’s well-known reduced gradient algorithm with the bundle method of nonsmooth optimization. The algorithm is explained clearly, convergence is proved, and five numerical examples are presented.
Reviewer: P.Loewen

##### MSC:
 90C25 Convex programming 65K05 Numerical mathematical programming methods 90C55 Methods of successive quadratic programming type
Full Text: