Numerical calculation of the pseudoinverse using Leonov's minimal method

Basic concepts and statements of the method of minimal pseudoinverse matrix developed by Leonov are considered. We prove that the minimal pseudoinverse matrix is a numerical ϵ- pseudoinverse and a comparative analysis with the method based on the singular value decomposition is presented. We also present a computational implementation coded using MATLAB language for the minimal method and we discuss the results of the numerical experimentation too.