News

Linear programming based analysis and solution of deterministic infinite horizon optimal control problems (discounting and time averaging cases)

ARGOMENTI: Convegni

SPECIAL SEMINAR FOR GRADUATE STUDENTS
Mercoledi' 12 marzo alle ore 12.15 in Aula 2AB/45 della Torre Archimede il professor Vladimir Gaitsgory (University South Australia) terra' un seminario dal titolo "Linear programming based analysis and solution of deterministic infinite horizon optimal control problems (discounting and time averaging cases)".

-Abstract
We will discuss relationships between infinite time horizon optimal control problems with discounting and time average criteria and certain infinite dimensional linear programming (IDLP) problems. We will state duality results for these IDLP problems and use these results for construction of necessary and sufficient optimality conditions and for characterization of viability kernels. We will also approximate IDLP problems and their duals by finite dimensional LP problems and their respective duals,
and we will use solutions of the latter for numerical construction of near optimal feedback controls in the corresponding optimal control problems. The presentation will be based on results included in [1], [2], [3].

-References
[1] V. Gaitsgory and S. Rossomakhine, "Linear Programming Approach to Deterministic Long Run Average Problems of Optimal Control", SIAM J. Control and Optimization, 44 (2005/2006), No 6, pp. 2006-2037.
[2] L. Finlay, V. Gaitsgory, and I. Lebedev, Duality In Linear Programming Problems Related to Long Run Average Problems of Optimal Control", SIAM J. Control and Optimization, Accepted for publication.
[3] V. Gaitsgory and M. Quincompoix, Linear programming analysis of deterministic infinite horizon optimal control problems (discounting and time averaging cases)", Submitted for publication.

Rif. int. M. Bardi, A. Cesaroni