Efficient numerical methods for entropy-linear programming problems (Q327229)

The subject of the article is located in a scientifically active area of entropy-linear programming (ELP). It usually includes a formulation into a maximization of entropy under affine constraints. The authors propose several numerical methods for solving ELP problems. They consider establishing sharp estimates for the convergence rates of the proposed methods. They show that the described algorithm can be applied to minimization problems for strongly convex functionals with affine functionals. The described approach is based on solving a specially Tikhonov regularized dual of an ELP problem. Explicit formulas are applied to the dual problem to obtain the solution of the original problem. They show that the solution of the original problem is with prescribed accuracy by using exact formulas deciding the number of fast gradient iterations which are necessary for the regularized dual problem. This article is well written, structured and explained, it contains five sections: Section 1: Introduction, Section 2: Ehrenfest model, Section 3: Entropy-linear programming problem, Section 4: Auxiliary results for the regularized dual of the ELP problem, Section 5: Main results, Section 6: Discussion of the main results, and Section 7: Concluding remarks. In fact, future scientific work on comparing the methods proposed in this paper with other numerical methods for ELP problems would be interesting.