Equivalent necessary and sufficient conditions on noise sequences for stochastic approximation algorithms
Keyword(s):
We consider stochastic approximation algorithms on a general Hilbert space, and study four conditions on noise sequences for their analysis: Kushner and Clark's condition, Chen's condition, a decomposition condition, and Kulkarni and Horn's condition. We discuss various properties of these conditions. In our main result we show that the four conditions are all equivalent, and are both necessary and sufficient for convergence of stochastic approximation algorithms under appropriate assumptions.
1996 ◽
Vol 28
(03)
◽
pp. 784-801
◽
2001 ◽
Vol 12
(04)
◽
pp. 533-550
◽
1980 ◽
Vol 35
(4)
◽
pp. 437-441
◽
1990 ◽
Vol 19
(1)
◽
pp. 31-37
◽
2015 ◽
Vol 719-720
◽
pp. 496-503
2018 ◽
Vol 16
(04)
◽
pp. 1850040
◽
2019 ◽
Vol 29
(08)
◽
pp. 1950110
◽
1996 ◽
Vol 53
(3)
◽
pp. 391-400
◽