Some examples of Bayesian adaptive programming*

https://doi.org/10.1016/S0019-9958(66)90138-0Get rights and content
Under an Elsevier user license
open archive

The Bayes learning in adaptive control processes is defined by the learning structure in which the unknown probability distribution is re-estimated a posteriori by the use of Bayes' theorem after the random variable is observed at each stage of the process. Three kinds of measures which evaluate the effect of the Bayes' learning are presented. Using three examples in which the adaptive control problem is completely solved, it is shown that Bayes learning is sometimes unreasonable, in a certain sense, if the length of the programming period is not large.

Cited by (0)

*

This work was supported in part by the Office of Naval Research (Grant NR 042-247).