Absolutely Expedient Algorithms

S. Lakshmivarahan
Additional contact information
S. Lakshmivarahan: University of Oklahoma, School of Electrical Engineering and Computer Science

Chapter Chapter 3 in Learning Algorithms Theory and Applications, 1981, pp 66-108 from Springer

Abstract: Abstract This chapter deals with the analysis and design of non-linear reward penalty learning algorithms of the absolutely expedient NAR−P type. A fundamental property that distinguishes this class of algorithms from those of Chapter 2 is that the set VM= εj | j = 1,2,…,M, ej = jth unit vector consisting of the corners of the simplex SM form the (only) absorbing states of the Markov process p(k). Each ej c VM is topologically closed and it will be shown that ej is also stochastically closed under NAR−P algorithms. Thus, in this case each element of v (there are M of them) according to problem 1.1 form an ergodic kernel as opposed to the only one ergodic kernel (which is SM) of the NER−P algorithms discussed in chapter 2. The presence of such multiple ergodic kernel makes the asymptotic behavior of the Markov process p(k) under this class of algorithms, as to be expected, very much dependent on the initial state as against the independence of the asymptotic behavior of p(k) on the initial state in case of the NER−P algorithms.

Keywords: Markov Process; Sample Path; Symmetry Condition; Regular Function; Learn Algorithm Theory (search for similar items in EconPapers)
Date: 1981
References: Add references at CitEc
Citations:

There are no downloads for this item, see the EconPapers FAQ for hints about obtaining it.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:spr:sprchp:978-1-4612-5975-6_3

Ordering information: This item can be ordered from
http://www.springer.com/9781461259756

DOI: 10.1007/978-1-4612-5975-6_3

Access Statistics for this chapter

More chapters in Springer Books from Springer
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().