|17 June, 2019|
|12:30 pm||a||1:30 pm|
Speaker: Miriam G. Báez Hernández (Veracruzana University of Mexico)
Title: “Approach schemes for some Infinite Linear Programming problems”
Date: Monday, July 17, 12:30 a.m.
Localication: CIO Seminar Room (Torretamarit Building)
Abstract. In this talk we will present approximation schemes for problems of Infinite Linear Programming. Some examples of problems brought to Infinite Linear Programming are: the Problem of Mass Transfer, Semi-Infinite Linear Programming, the Transboundary Problem and the Markov Control Problem. One of the most important techniques of Infinite Linear Programming is the approach theory, for which it is necessary to look for conditions under which there are solutions for a particular problem. Hernández-Lerma and Lasserre propose a general approximation scheme for infinite linear programs, which requires two procedures: aggregation-relaxation of the constraints and internal approximation of the variable of interest. Applying the above, the Infinite Linear Programming problem is discretized. In this talk, we will show the computational implementation of the scheme proposed by Hernández-Lerma and Lasserre, to the Problem of Mass Transfer of Monge-Kantorovich. We will end the talk with an approximation scheme for the Markov Control Problem with discount criteria. Hernández-Lerma and Lasserre prove that the Markov Control Problem with discount criteria is equivalent to an Infinite Linear Programming problem when using occupation measures, using the equivalent linear programming problem and under the proposed algorithm for the Transfer Problem of Masses of Monge-Kantorovich proposed by Gabriel-Argüelles, López-Martínez and González-Hernádez we will show an approximation scheme in compact spaces.