Strong convergence of estimators as ε n -minimisers of optimisation problemsof optimisation problems

In the paper we prove strong consistency of estimators as solution of optimisation problems. The approach of the paper covers non-identifiable models, and models for dependent samples. We provide statements about consistency of M-estimators in regression models with random and with non-random design. The research was partially supported by the Deutsche Forschungsgemeinschaft (project number 436TSE113/40) and by the Grant Agency of the Czech Republic under Grant 201/03/1027.