Approximate Dynamic Programming Based on High Dimensional Model Representation

0399560 - ÚTIA 2014 RIV CZ eng J - Journal Article
Pištěk, Miroslav
Approximate Dynamic Programming Based on High Dimensional Model Representation.
Kybernetika. Roč. 49, č. 5 (2013), s. 720-737. ISSN 0023-5954
R&D Projects: GA ČR(CZ) GAP102/11/0437
Institutional support: RVO:67985556
Keywords : approximate dynamic programming * Bellman equation * approximate HDMR minimization * trust region problem
Subject RIV: BC - Control Systems Theory
Impact factor: 0.563, year: 2013
http://library.utia.cas.cz/separaty/2013/AS/pistek-0399560.pdf

This article introduces an algorithm for implicit High Dimensional Model Representation (HDMR) of the Bellman equation. This approximation technique reduces memory demands of the algorithm considerably. Moreover, we show that HDMR enables fast approximate min- imization which is essential for evaluation of the Bellman function. In each time step, the problem of parametrized HDMR minimization is relaxed into trust region problems, all sharing the same matrix. Finding its eigenvalue decomposition, we effectively achieve estimates of all minima. Their full-domain representation is avoided by HDMR and then the same approach is used recursively in the next time step. An illustrative example of N-armed bandit problem is included. We assume that the newly established connection between approximate HDMR minimization and the trust region problem can be beneficial also to many other applications.
Permanent Link: http://hdl.handle.net/11104/0226953