Institutional Repository
Technical University of Crete
Technical University of Crete
EN
|
EL
| URI: | http://purl.tuc.gr/dl/dias/4CBA1F85-07E4-4B31-AE63-9C5FE9ADAFF8 | ||
| Year | 1980 | ||
| Type of Item | Peer-Reviewed Journal Publication | ||
| License |
|
||
| Bibliographic Citation | A.D. Pouliezos ," An iterative method for calculating the sample serial correlation coefficient ," IEEE Trans. on Aut. Control ,vol.25,no.4 ,pp.:834 - 836,1980.doi:10.1109/TAC.1980.1102405 https://doi.org/10.1109/TAC.1980.1102405 | ||
| Appears in Collections |
Generally, when stochastic approximation is used to identifythe ooefficients of a linear system or for an adaptive fUter or equalizer, theiterate Xn is projected back onto some finite set G= {x: lx;l< B, all i}, ifit ever leaves it. The convergence of such truncated sequences have beendiscussed Informally. Here it is shown, under very broad conditions on thenoises, that {Xn} converges with probabUity 1 to the closest point in G tothe optimum value of Xn. Also, under even weaker conditions, the case ofconstant coefficient sequence is treated and a weak convergence resultobtained. The set G is used for simpHcity. It can be seen that the resultholds true in more general cases, but the box is used since it is the onlycommonly used constraint set for this problem.
Copyright © DIAS 2013
