Regularization methods for unstable minimization problems based on the idea of set extension (Q752011)

An optimization problem is said to be stable, or well posed, if for a small perturbation of the coefficients we obtain small variations of the solutions. Unstable or `ill-posed' problems need regularizing treatment to make them more stable. In this paper three regularization methods are studied, based on the idea of an extension of the feasible domain and estimations of the speed of convergence, with respect to the problem's functions and argument, are given for all three methods.