Probabilistic Programming with Applications

Titulo:  «Probabilistic Programming with Applications»

Ponente: René Henrion (Weierstrass Institute for Applied Analysis and Stochastics)

Fecha: Jueves 13 de septiembre de 2018, 12:30 h.

Lugar: Aula 0.2, Instituto Universitario de Investigación CIO, Edificio Torretamarit, Universidad Miguel Hernández (Campus de Elche)

Resumen:  Optimization problems with probabilistic constraints (or chance constraints) represent a major model of stochastic programming with numerous applications in engineering. Basically, in the presence of random parameters, a decision is declared to be feasible if it satisfies a given system of random inequalities with a prescribed high probability (e.g., 90%). The mathematical challenge of such problems is the absence of explicit formulae for the probability as a function of the decision. This makes the structural analysis (e.g., (semi-) continuity, differentiability, convexity etc.) and the algorithmic solution much more difficult than in a conventional setting. The talk provides some ideas in this direction along with applications to energy management.