Institute of Information Theory and Automation

You are here

Bibliography

Journal Article

Fully probabilistic control design in an adaptive critic framework

Herzallah R., Kárný Miroslav

: Neural Networks vol.24, 10 (2011), p. 1128-1135

: CEZ:AV0Z10750506

: GA102/08/0567, GA ČR

: Stochastic control design, Fully probabilistic design, Adaptive control, Adaptive critic

: 10.1016/j.neunet.2011.06.006

: http://library.utia.cas.cz/separaty/2011/AS/karny-0364820.pdf

(eng): Optimal stochastic controller pushes the closed-loop behavior as close as possible to the desired one. The fully probabilistic design (FPD) uses probabilistic description of the desired closed loop and minimizes Kullback–Leibler divergence of the closed-loop description to the desired one. Practical exploitation of the fully probabilistic design control theory continues to be hindered by the computational complexities involved in numerically solving the associated stochastic dynamic programming problem; in particular, very hard multivariate integration and an approximate interpolation of the involved multivariate functions. This paper proposes a new fully probabilistic control algorithm that uses the adaptive critic methods to circumvent the need for explicitly evaluating the optimal value function, thereby dramatically reducing computational requirements. This is a main contribution of this paper.

: BC

2019-01-07 08:39