I am reading the book Introduction to Reinforcement Learning by Barto and Sutton.
Snippet from book:
What is this well-known result/law used here?
I am reading the book Introduction to Reinforcement Learning by Barto and Sutton.
Snippet from book:
What is this well-known result/law used here?
Copyright © 2021 JogjaFile Inc.

This result is mentioned in the Wikipedia article for stochastic approximation in the section on the Robbins-Munro algorithm, although this result doe not appear in (RM51). It does appear explicitly in (B54). In the survey (D56), these conditions are called "Blum's conditions", but I don't see evidence in other literature of use of that phrase or any other short phrase to describe the set of conditions you quote.
(B54): Blum, Julius R., "Approximation Methods which Converge with Probability one", Ann. Math. Stat. 25 (2): 382–386, 01 June 1954. doi:10.1214/aoms/1177728794.
(D56): Derman, C., "Stochastic Approximation", Ann. Math. Stat. 27 (4): 879-886, 1956. doi:10.1214/aoms/1177728065
(RM51): Robbins, H.; Monro, S., "A Stochastic Approximation Method", Ann. Math. Stat. 22 (3): 400, 1951. doi:10.1214/aoms/1177729586.