Sur la convergence d’une méthode associant penalisation et regularisation. (On the convergence of a method combining penalization and regularization). (French) Zbl 0641.90065

Mouallif, Khalil

Bull. Soc. R. Sci. Liège 56, 175-180 (1987).

Summary: A method for solving a general constrained nonsmooth convex optimization problem is considered, which combines penalty methods with the proximal method. Some conditions are given under which, the whole sequence generated by such a method, converges to an optimal point. An application is made to convex programming in a paper by the author and P. Tossings [ibid. 56, 181-190 (1987; Zbl 0623.90062)].

MSC:

90C25	Convex programming
65K05	Numerical mathematical programming methods

Keywords:

convergence analysis; constrained nonsmooth convex optimization; penalty methods; proximal method

Citations:

Zbl 0623.90062