000196322 001__ 196322
000196322 005__ 20190812205738.0
000196322 037__ $$aCONF
000196322 245__ $$aRobust Data-Driven Dynamic Programming
000196322 269__ $$a2013
000196322 260__ $$c2013
000196322 336__ $$aConference Papers
000196322 520__ $$aIn stochastic optimal control the distribution of the exogenous noise is typically unknown and must be inferred from limited data before dynamic programming (DP)-based solution schemes can be applied. If the conditional expectations in the DP recursions are estimated via kernel regression, however, the historical sample paths enter the solution procedure directly as they determine the evaluation points of the cost-to-go functions. The resulting data-driven DP scheme is asymptotically consistent and admits efficient computational solution when combined with parametric value function approximations. If training data is sparse, however, the estimated cost-to-go functions display a high variability and an optimistic bias, while the corresponding control policies perform poorly in out-of-sample tests. To mitigate these small sample effects, we propose a robust data-driven DP scheme, which replaces the expectations in the DP recursions with worst-case expectations over a set of distributions close to the best estimate. We show that the arising min-max problems in the DP recursions reduce to tractable conic programs. We also demonstrate that this robust algorithm dominates state-of-the-art benchmark algorithms in out-of-sample tests across several application domains.
000196322 700__ $$aHanasusanto, Grani A.
000196322 700__ $$g239987$$aKuhn, Daniel$$0247589
000196322 7112_ $$dDecember 2013$$cLake Tahoe, USA$$aNeural Information Processing Systems
000196322 720_1 $$aBurges, C. J. C.$$eed.
000196322 720_1 $$aBottou, L.$$eed.
000196322 720_1 $$aWelling, M.$$eed.
000196322 720_1 $$aGhahramani, Z.$$eed.
000196322 720_1 $$aWeinberger, K. Q.$$eed.
000196322 773__ $$tNIPS Proceedings 26
000196322 8564_ $$zURL$$uhttp://papers.nips.cc/paper/5123-robust-data-driven-dynamic-programming
000196322 8560_ $$fdaniel.kuhn@epfl.ch
000196322 85641 $$uhttp://papers.nips.cc/paper/5123-robust-data-driven-dynamic-programming.pdf
000196322 909C0 $$xU12788$$pRAO$$0252496
000196322 909CO $$ooai:infoscience.tind.io:196322$$qGLOBAL_SET$$pconf$$pCDM
000196322 917Z8 $$x112541
000196322 937__ $$aEPFL-CONF-196322
000196322 973__ $$rREVIEWED$$aEPFL
000196322 980__ $$aCONF