Learning effective state-feedback controllers through efficient multilevel importance samplers
| Autor: | S. A. Menchón, Hilbert J. Kappen |
|---|---|
| Rok vydání: | 2018 |
| Předmět: |
0209 industrial biotechnology
Computer science Ciencias Físicas Control (management) Biophysics MathematicsofComputing_NUMERICALANALYSIS 02 engineering and technology Nonlinear control Otras Ciencias Físicas MULTILEVEL MONTE CARLO METHOD purl.org/becyt/ford/1 [https] 020901 industrial engineering & automation Control theory IMPORTANCE SAMPLING 0202 electrical engineering electronic engineering information engineering Class (computer programming) Sampling (statistics) purl.org/becyt/ford/1.3 [https] Computer Science Applications PATH INTEGRAL CONTROL PROBLEMS Control and Systems Engineering Path integral formulation 020201 artificial intelligence & image processing State (computer science) CIENCIAS NATURALES Y EXACTAS Importance sampling |
| Zdroj: | International Journal of Control, 92, 12, pp. 2776-2783 International Journal of Control, 92, 2776-2783 CONICET Digital (CONICET) Consejo Nacional de Investigaciones Científicas y Técnicas instacron:CONICET |
| ISSN: | 1366-5820 0020-7179 |
| Popis: | Monte Carlo sampling can be used to estimate the solution of path integral control problems, which are a restricted class of nonlinear control problems with arbitrary dynamics and state cost, but with a linear dependence of the control on the dynamics and quadratic control cost. Although importance sampling is used to improve numerical computations, the effective sample size may still be low or many samples could be required. In this work, we propose a method to learn effective state-feedback controllers for nonlinear stochastic control problems based on multilevel importance samplers. In particular, we focus on the question of how to compute effective importance samplers considering a multigrid scenario. We test our algorithm in finite horizon control problems based on Lorenz-96 model with chaotic and non-chaotic behaviour, showing, in all cases, that our multigrid implementation reduces the computational time and improves the effective sample size. Fil: Menchón, Silvia Adriana. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Córdoba. Instituto de Física Enrique Gaviola. Universidad Nacional de Córdoba. Instituto de Física Enrique Gaviola; Argentina. Universidad Nacional de Córdoba. Facultad de Matemática, Astronomía y Física; Argentina. Radboud Universiteit Nijmegen. Donders Instituto Brain Cognition and Behavior. SNN Machine Learning Group; Países Bajos Fil: Kappen, Hilbert Johan. Radboud Universiteit Nijmegen. Donders Instituto Brain Cognition and Behavior. SNN Machine Learning Group; Países Bajos |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|