• Aucun résultat trouvé

Cross-entropy optimization of control policies with adaptive basis functions

N/A
N/A
Protected

Academic year: 2021

Partager "Cross-entropy optimization of control policies with adaptive basis functions"

Copied!
14
0
0

Texte intégral

Loading

Figure

Fig. 1. Schematic of the policy parameterization. The vector ϑ associates the BFs to the discrete actions
Fig. 2. Optimal policy for the double integrator. Black corresponds to the action − 0.1, and white corresponds to + 0.1.
Fig. 3. Performance of CE policy search.
Fig. 7. Performance of DIRECT—comparison with CE optimization.
+4

Références

Documents relatifs

As explained next, the illumination problem (2) is somewhat related with the problem of finding the circumcenter of a polyhedral convex

Energy-optimal drive control under several scenarios (acceleration, decelera- tion, driving between stops, approaching traffic lights, cruise within a speed band) [3] can be treated

Bergounioux, Optimal control of abstract elliptic variational inequalities with state constraints,, to appear, SIAM Journal on Control and Optimization, Vol. Tiba, General

We first penalize the state equation and we obtain an optimization problem with non convex constraints, that we are going to study with mathematical programming methods once again

Index Terms—Stochastic Dynamic Programming, Policy Iteration Algorithm, Autoregressive Models, Ocean Wave Energy, Power Smoothing.. 1 I NTRODUCTION TO P OWER

As indicated above, the homogenization method and the classical tools of non-convex variational problems (in particular, Young measures) are, for the moment, two of the most

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des